Engine testing theory and practice

Engine Testing

This page intentionally left blank

Engine TestingTheory and Practice

Third edition

A.J. MartyrM.A. Plint

AMSTERDAM • BOSTON • HEIDELBERG • LONDON • NEWYORK • OXFORD

PARIS • SANDIEGO • SAN FRANCISCO • SINGAPORE • SYDNEY • TOKYO

Butterworth-Heinemann is an imprint of Elsevier

Butterworth-Heinemann is an imprint of ElsevierLinacre House, Jordan Hill, Oxford OX2 8DP30 Corporate Drive, Suite 400, Burlington, MA 01803

First edition 1995Reprinted 1996 (twice), 1997 (twice)Second edition 1999Reprinted 2001, 2002Third edition 2007

Copyright © 2007, A.J. Martyr and M.A. Plint. Published by Elsevier Ltd. All rights reserved

The right of A.J. Martyr and M.A. Plint to be identified as the authors of this work has been assertedin accordance with the Copyright, Designs and Patents Act 1988

No part of this publication may be reproduced, stored in a retrieval system, ortransmitted in any form or by any means electronic, mechanical, photocopying,recording or otherwise without the prior written permission of the publisher

Permissions may be sought directly from Elsevier’s Science & Technology RightsDepartment in Oxford, UK: phone (+44) (0) 1865 843830; fax: (+44) (0) 1865 853333;email: [email protected]. Alternatively you can submit your request online byvisiting the Elsevier web site at http://elsevier.com/locate/permissions, and selectingObtaining permission to use Elsevier material

NoticeNo responsibility is assumed by the publisher for any injury and/or damage to personsor property as a matter of products liability, negligence or otherwise, or from any useor operation of any methods, products, instructions or ideas contained in the materialherein. Because of rapid advances in the medical sciences, in particular, independentverification of diagnoses and drug dosages should be made

British Library Cataloguing in Publication Data

A catalogue record for this book is available from the British Library

Library of Congress Cataloging-in-Publication Data

A catalog record for this book is available from the Library of Congress

ISBN-13: 978-0-7506-8439-2

For information on all Butterworth-Heinemann publicationsvisit our web site at http://books.elsevier.com

Printed and bound in the UK

07 08 09 10 10 9 8 7 6 5 4 3 2 1

Working together to grow libraries in developing countries

www.elsevier.com | www.bookaid.org | www.sabre.org

Contents

Preface vii

Acknowledgements ix

Introduction xi

Units and conversion factors xv

1 Test facility specification, system integration and project organization 1

2 The test cell as a thermodynamic system 14

3 Vibration and noise 21

4 Test cell and control room design: an overall view 47

5 Ventilation and air conditioning 72

6 Test cell cooling water and exhaust gas systems 108

7 Fuel and oil storage, supply and treatment 129

8 Dynamometers and the measurement of torque 144

9 Coupling the engine to the dynamometer 170

10 Electrical design considerations 197

11 Test cell control and data acquisition 216

12 Measurement of fuel, combustion air and oil consumption 242

13 Thermal efficiency, measurement of heat and mechanical losses 263

14 The combustion process and combustion analysis 282

15 The test department organization, health and safety management, riskassessment correlation of results and design of experiments 308

16 Exhaust emissions 324

17 Tribology, fuel and lubrication testing 354

18 Chassis or rolling road dynamometers 368

19 Data collection, handling, post-test processing, engine calibrationand mapping 395

20 The pursuit and definition of accuracy: statistical analysis of test results 408

Index 423


Preface

The preface of this book is probably the least read section of all; however, it isthe only part in which I can pay tribute to my friend and co-author of the first twoeditions, Dr Michael Plint, who died suddenly in November 1998, only four daysafter the publication of the second edition.

All the work done by Michael in the previous editions has stood up to the scrutinyof our readers and my own subsequent experience. In this edition, I have attemptedto bring our work up to date by revising the content to cover the changing legislation,techniques and some of the new tools of our industry. In a new Chapter 1, I have alsosought to suggest some good practices, based on my own 40 years of experience,aimed at minimizing the problems of project organization that are faced by all partiesinvolved in the specification, modification, building and commissioning of enginetest laboratories.

The product of an engine test facility is data and byproduct is the experiencegained by the staff and hopefully retained by the company. These data have tobe relevant to the experiments being run, and every component of the test facilityhas to play its part, within an integrated whole, in ensuring that the test data areas valid and uncorrupted as possible, within the sensible limits of the facility’srole. It was our intention when producing the first edition to create an eclecticsource of information that would assist any engineer faced with the many designand operational problems of both engine testing and engine test facilities. In theintervening years, the problems have become more difficult as the nature of theengine control has changed significantly, while the time and legislative pressureshave increased. However, it is the laws of physics that rule supreme in our worldand they can continue to cause problems in areas outside the specialization of manyindividual readers. I hope that this third edition helps the readers involved in someaspect of engine testing to gain a holistic view of the whole interactive package thatmakes up a test facility and to avoid, or solve, some of the problems that they maymeet in our industry.

Having spoken to a number of readers of the two proceeding editions of this bookI have reorganized the contents of most of the chapters in order to reflect the way inwhich the book is used.

Writing this edition has, at times, been a lonely and wearisome task that wouldnot have been completed without the support of my wife Diana and my friends.Many people have assisted me with their expert advice in the task of writing thisthird edition. I have to thank all my present AVL colleagues in the UK and Austria,particularly Stuart Brown, David Moore and Colin Freeman who have shared many

viii Preface

of my experiences in the test industry over the last 20 years, also Dave Rogers, CraigAndrews, Hans Erlach and finally Gerhard Müller for his invaluable help with thecomplexities of electrical distribution circuits.

A.J. MartyrInkberrow

22 September 2006

Acknowledgements

Figures 3.6, 3.13, 3.15 and 3.16 Reprinted from Industrial Noise Control, Fader, bypermission of John Wiley and Sons LtdFigure 3.9 Reprinted from technical literature of Type TSC by courtesy of Christieand Grey Ltd, UKFigure 3.14 Reprinted from Encyclopaedia of Science and Technology, Vol. 12,1987, by kind permission of McGraw-Hill Inc., New YorkFigure 5.3 Reprinted from CIBSE Guide C, section C4, by permission of the Char-tered Institution of Building Services EngineersFigure 5.10 Reprinted from I.H.V.E. Psychometric Chart, by permission of theChartered Institution of Building Services EngineersFigure 7.1 Reprinted from BS799. Extracts from British Standards are reprinted withthe permission of BSI. Complete copies can be obtained by post from BSI Sales,Linford Wood, Milton Keynes, MK14 6LE, UKFigure 7.2 Reprinted from The Storage and Handling of Petroleum Liquids, Hughesand Swindells, with the kind permission of Edward Arnold PublishersFigure 7.3 Reprinted from ‘Recommendations for pre-treatment and cleaning ofheavy fuel oil’ with the kind permission of Alfa Laval LtdFigure 8.3 Reprinted from Drawing no. GP10409 (Carl Shenck AG, Germany)Figures 8.4 and 8.5 Reprinted from Technical Documentation T 32 FN, with thekind permission of Hottinger Baldwin Messtechnik GmbH, GermanyFigures 8.6 and 8.7 Illustration courtesy of Ricardo Test Automation LtdFigures 8.9, 8.10 and 8.11 Reprinted with permission of Froude Consine, UKFigure 8.12 Reprinted from technical literature, Wichita LtdFigure 9.4 Reprinted from Practical Solution of Torsional Vibration Problems, 3rdedition, W. Ker-Wilson, 1956Figure 9.7 Reprinted from literature, with the kind permission of British AutoguardLtdFigures 9.8 and 9.11 Reprinted from sales and technical literature, with the kindpermission of Twilfex LtdFigure 12.2 Reprinted from Paper ISATA, 1982, R.A. Haslett, with the kind permis-sion of Cussons LtdFigure 12.4 Reprinted from Technology News with the kind permission of PetroleumReviewFigure 14.7 Reprinted from SAE 920 462 (SAE International Ltd)Figure 17.1 From Schmiertechnik und Tribologie 29, H. 3, 1982, p. 91, VincentVerlag Hannover (now: Tribologie und Schmierungstechnik)

x Acknowledgements

Figures 17.2 and 17.3 Reprinted by permission of the Council of the Institutionof Mechanical Engineers from ‘The effect of viscosity grade on piston ring wear’,S.L. Moore, Proc. I. Mech. E C184/87Figure 18.2 Illustration courtesy of Ricardo Test Automation LtdFigures 2.3, 5.4, 5.5, 5.6, 5.7, 6.8, 7.5, 7.6, 10.5 10.8, 14.9, 14.10, 14.11, 16.1, 16.2,

16.5, 16.7, 16.8 and 18.4 Reprinted by kind permission of AVL List GmbH

Introduction

Over the working lifetime of the authors the subject of internal engine developmentand testing has changed, from being predominantly within the remit of mechanicalengineers, into a task that is well beyond the remit of any one discipline that requiresa team of specialists covering, in addition to mechanical engineering, electronics,power electrics, acoustics, software, computer sciences and chemical analysis, allsupported by expertise in building services and diverse legislation.

It follows that the engineer concerned with any aspect of engine testing, be itfundamental research, development, performance monitoring or routine productiontesting, must have at his fingertips a wide and ever-broadening range of knowledgeand skills.

A particular problem he must face is that, while he is required to master ever moreadvanced experimental techniques – such areas as emissions analysis and enginecalibration come to mind – he cannot afford to neglect any of the more traditionalaspects of the subject. Such basic matters as the mounting of the engine, coupling itto the dynamometer and leading away the exhaust gases can give rise to intractableproblems, misleading results and even on occasion to disastrous accidents. More thanone engineer has been killed as a result of faulty installation of engines on test beds.

The sheer range of machines covered by the general term internal combustion

engine broadens the range of necessary skills. At one extreme we may be concernedwith an engine for a chain saw, a single cylinder of perhaps 50 c.c. capacity runningat 15 000 rev/min on gasoline, with a running life of a few hours. Then we havethe vast number of passenger vehicle engines, four, six or eight cylinder, capacitiesranging from one litre to six, expected to develop full torque over speeds rangingfrom perhaps 1500 rev/min up to 7000 rev/min (the upper limit rising continually),and with an expected life of perhaps 6000 hours. The motor-sport industry continuesto push the limits of both engine and test plant design with engines revving atspeeds approaching 20 000 r.p.m. and, in rally cars, engine control systems havingto cope with cars leaving the ground, then requiring full power when they land. Atthe other extreme is the cathedral type marine engine, a machine perhaps 10m talland weighing 1000 tonnes, running on the worst type of residual fuel, and expectedto go on turning at 70 rev/min for more than 50 000 hours.

The purpose of this book is to bring together the information on both the theoryand practice of engine testing that any engineer responsible for work of this kindmust have available. It is naturally not possible, in a volume of manageable size, togive all the information that may be required in the pursuit of specialized lines ofdevelopment, but it is the intention of the authors to make readers aware of the many

xii Introduction

tasks they may face and to give advice based on experience; a range of referencesfor more advanced study has been included.

Throughout the book accuracy will be a recurring theme. The purpose of enginetesting is to produce information, and inaccurate information can be useless or worse.A feeling for accuracy is the most difficult and subtle of all the skills required of thetest engineer. Chapter 19, dealing with this subject, is perhaps the most important inthe book and the first that should be read.

Experience in the collaboration with architects and structural engineers is par-ticularly necessary for engineers involved in test facility design. These professionsfollow design conventions and even draughting practices that differ from those ofthe mechanical engineer. To give an example, the test cell designer may specify astrong floor on which to bolt down engines and dynamometers that has an accuracyapproaching that of a surface plate. To the structural engineer this will be a startlingconcept, not easily achieved.

The internal combustion engine is perhaps the best mechanical device availablefor introducing the engineering student to the practical aspects of engineering. Anengine is a comparatively complicated machine, sometimes noisy and alarming in itsbehaviour and capable of presenting many puzzling problems and mystifying faults.A few hours spent in the engine testing laboratory are perhaps the best possibleintroduction to the real world of engineering, which is remote from the world of thelecture theatre and the computer simulation in which, inevitably, the student spendsmuch of his time.

While it contains some material only of interest to the practising test engineer,much of this book is equally suitable as a student text, and this purpose has been keptvery much in mind by the authors. In response to the author’s recent experience, thethird edition has a new Chapter 1 dedicated to the problems involved in specifyingand managing a test facility build project.

A note of warning: the general management of engine tests

What may be regarded as traditional internal combustion engines had in generalvery simple control systems. The spark ignition engine was fitted with a carburettorcontrolled by a single lever, the position of which, together with the resisting torqueapplied to the crankshaft, set all the parameters of engine operation. Similarly, theperformance of a diesel engine was dictated by the position of the fuel pump rack,either controlled directly or by a relatively simple speed governor.

The advent of engine control units (ECUs) containing ever more complex mapsand taking signals from multiple vehicle transducers has entirely changed the sit-uation. The ECU monitors many aspects of powertrain performance and makescontinuous adjustments. The effect of this is effectively to take the control of thetest conditions out of the hands of the engineer conducting the test. Factors entirelyextraneous to the investigation in hand may thus come into play.

Introduction xiii

The introduction of exhaust gas recirculation (EGR) under the control of the ECUis a typical example. The only way open to the test engineer to regain control of histest is to devise means of bypassing the ECU, either mechanically or by interventionin the programming of the control unit.

A note on references and further information

It would clearly not be possible to give all the information necessary for the practiceof engine testing and the design of test facilities in a book of this length. Referencessuitable for further study are given at the end of most chapters. These are of twodifferent kinds:

• a selection of fundamental texts or key papers• relevant British Standards and other reference standard specifications.

The default source of many students is now the world wide web which containsvast quantities of information related to engines and engine testing, much of whichis written by and for the automotive after-market where a rigorous approach toexperimental accuracy is not always evident; for this reason and due to the transientnature of many websites, there are very few web-based references.


Units and conversion factors

Throughout this book use is made of the metric system of units, variouslydescribed as:

The MKS (metre-kilogram-second) SystemSI (Système International) Units

These units have the great advantage of logical consistency but the disadvantages ofstill a certain degree of unfamiliarity and in some cases of inconvenient numericalvalues.

Fundamental UnitsMass kilogram (kg) 1 kg= 2.205 lbLength metre (m) 1m= 39.37 inForce newton (N) 1N= 0.2248 lbf

Derived UnitsArea square metre (m2) 1m2 = 10.764 ft2

Volume cubic metre (m3), litre (l) 1m3 = 10001= 35.3 ft3

Velocity metre per second (m/s) 1m/s= 3.281 ft/sWork, Energy joule (J) 1 J= 1Nm= 0.7376 ft-lbfPower watt (W) 1W= 1 J/s

1 horsepower (hp)= 745.7WTorque newton metre 1Nm= 0.7376 lbf-ft

The old metric unit of energy was the calorie (cal), the heat to raise the temperatureof 1 gram of water by 1�C.

1 cal= 4.1868 J. 1 kilocalorie (kcal) = 4.1868 kJ.

Temperature degree Celsius �C �

Absolute temperature kelvin K T

T = �+273�15

Pressure pascal (Pa) 1 Pa= 1N/m2 = 1.450× 10−4 lbf/in2

1MPa= 106 Pa= 145 lbf/in2

xvi Units and conversion factors

This unit is commonly used to denominate stress.Throughout this book the bar is used to denominate pressures:

1 bar (bar)= 105 Pa= 14.5 lbf/in3

Standard test conditions for i.c. engines as defined in BS 5514/ISO 30461 specify:

Standard atmospheric pressure = 1 bar = 14.5 lbf/in2

Note: ‘Standard atmosphere’ as defined by the physicist2 is specified as a barometricpressure of 760 millimetres of mercury (mmHg) at 0�C.

1 standard atmosphere = 1.01325 bar = 14.69 lbf/in2

The difference between these two standard pressures is a little over 1 per cent. This

can cause confusion. Throughout this book 1 bar is regarded as standard atmospheric

pressure.

The torr is occasionally encountered in vacuum engineering.

1 torr = 1mmHg = 133.32 Pa

In measurements of air flow use is often made of water manometers.

1mm of water (mmH2O) = 9.81 Pa

References

1. BS 5514 Reciprocating Internal Combustion Engines: Performance.

2. Kaye, G.W.C. and Laby, T.H. (1973) Tables of Physical and Chemical Constants,Longmans, London.

Further reading

BS 350 Pt 1 Conversion factors and tables

BS 5555 Specification for SI units and recommendations for the use of their multiplesand of certain other units

1 Test facility specification,system integration and projectorganization

Introduction

An engine test facility is a complex of machinery, instrumentation and supportservices, housed in a building adapted or built for its purpose. For such a facilityto function correctly and cost-effectively, its many parts must be matched to eachother while meeting the operational requirements of the user and being compliantwith various regulations.

Engine and vehicle developers now need to measure improvements in engine per-formance that are frequently so small as to require the best available instrumentationin order for fine comparative changes in performance to be observed. This level ofmeasurement requires that instrumentation is integrated within the total facility suchthat their performance and data are not compromised by the environment in whichthey operate and services to which they are connected.

Engine test facilities vary considerably in power rating and performance; in addi-tion there are many cells designed for specialist interests, such as production test orstudy of engine noise, lubrication oils or exhaust emissions.

The common product of all these cells is data that will be used to identify, modify,homologate or develop performance criteria of all or part of the tested engine. Allpost-test work will rely on the relevance and veracity of the test data, which in turnwill rely on the instrumentation chosen to produce it and the system within whichthe instruments work.

To build or substantially modify a modern engine test facility requires co-ordination of a wide range of specialized engineering skills; many technical managershave found it to be an unexpectedly complex task.

The skills required for the task of putting together test cell systems from theirmany component parts have given rise, particularly in the USA, to a specializedindustrial role known as system integration. In this industrial model, a company ormore rarely a consultant, having one of the core skills required, takes contractualresponsibility for the integration of all of the test facility components from varioussources. Commonly, the integrator role has been carried out by the supplier of test cellcontrol systems and the role has been restricted to the integration of the dynamometerand control room instrumentation.

2 Engine Testing

In Europe, the model is somewhat different because of the long-term developmentof a dynamometry industry that has given rise to a very few large test plant contractingcompanies.

However, the concept of systems integrator is useful to define that role, within aproject, that takes the responsibility for the final functionality of a test facility; sothe term will be used, where appropriate, in the following text.

This chapter covers the vital importance of good user specification and the variousorganizational structures required to complete a successful test facility project.

Test facility specification

Without a clear and unambiguous specification no complex project should be allowedto proceed.

This book suggests the use of three levels of specification:

1. Operational specification: describing ‘what it is for’, created by the user priorto any contract to design or build a test facility.

2. Functional specification: describing ‘what it consists of andwhere it goes’, createdeither by the user group having the necessary skills, as part of a feasibility study bya third party, or by the main contractor as part of the first phase of a contract.

3. Detailed functional specification: describing ‘how it all works’ created by theproject design authority within the supply contract.

Creation of an operational specification

This chapter will tend to concentrate on the operational specification which is a user-generated document, leaving some aspects of the more detailed levels of functionalspecification to subsequent chapters covering the design process. The operational spec-ification should contain a clear description of the task for which the facility is beingcreated. It need not specify in detail the instruments required, nor does it have to bebased on a particular site. The operational specification is produced by the end user;its first role will normally be to support the application for budgetary support and out-line planning; subsequently, it remains the core document on which all other detailedspecifications are based. It is sensible to include a brief description of envisaged facilityacceptance tests within the document since there is no better means of developing andcommunicating the user’s requirement than to describe the results to be expected fromdescribed work tasks.

• It is always sound policy to find out what is available on the market at an earlystage, and to reconsider carefully any part of the specification that makes demandsthat exceed what is commonly offered.

• A general cost consciousness at this stage can have a permanent effect on capitaland subsequent running costs.

Test facility specification, system integration and project organization 3

Because of the range of skills required in the design and commissioning of a ‘greenfield’ test laboratory it is remarkably difficult to produce a succinct specification thatis entirely satisfactory, or even mutually comprehensible, to all specialist participants.

The difficulty is compounded by the need for some of the building design detailsthat determine the final shape, such as roof penetrations or floor loadings, to be deter-mined before the detailed design of internal plant has been finalized. It is appropriatethat the operational specification document contains statements concerning the gen-eral ‘look and feel’ and any such pre-existing conditions or imposed restrictions thatmay impact on the facility layout. It should list any prescribed or existing equipmentthat has to be integrated, the level of staffing and any special industrial standardsthe facility is required to meet. In summary, it should at least address the followingquestions:

• What are the primary and secondary purposes for which the facility is intendedand can these functions be condensed into a sensible set of acceptance proceduresto prove the purposes that may be achieved?

• What is the realistic range of units under test (UUT)?• How are test data (the product of the facility) to be displayed, distributed, stored

and post-processed?• What possible extension of specification or further purposes should be provided

for in the initial design and to what extent would such ‘future proofing’ distortthe project phase costs?

• May there be a future requirement to install additional equipment and how willthis affect space requirement?

• Where will the UUT be prepared for test?• How often will the UUT be changed and what arrangements will be made for

transport into and from the cells?• How many different fuels are required and must arrangements be made for

quantities of special or reference fuels?• What up-rating, if any, will be required of the site electrical supply and distribution

system?• To what degree must engine vibration and exhaust noise be attenuated within the

building and at the property border?• Have all local regulations (fire, safety, environment, working practices, etc.) been

studied and considered within the specification?

Feasibility studies and outline planning permission

The work required to produce a site-specific operational specification, or statementof intent, may produce a number of alternative layouts each with possible first-cost or operational problems. In all cases an environmental impact report should beproduced covering both the facility’s impact of its surroundings and, in the case oflow emission measuring laboratories, the locality’s impact on the facility.

4 Engine Testing

Complex technocommercial investigatory work may be needed so a feasibilitystudy might be considered, covering the total planned facility or that part that givesrise to doubt or the subject of radically differing strategies. In the USA, this type ofwork is often referred to as a proof design contract.

The secret of success of such studies is the correct definition of the required‘deliverable’ which must answer the technical and budgetary dilemmas, give clearand costed recommendations and, so far as is possible, be supplier neutral. Thefinal text should be capable of easy incorporation into the Operation and FunctionalSpecification documents.

A feasibility study will invariably be concerned with a specific site and, providingappropriate expertise is used, should prove supportive to gaining budgetary andoutline planning permission; to that end, it should include within its content a sitelayout drawing and graphical representation of the final building works.

Benchmarking

Cross-referencing with other test facilities or test procedures is always useful whenspecifying your own. Benchmarking is merely a modern term for an activity thathas been practised by makers of products intended for sale, probably ever since thefirst maker of flint axes went into business: it is the act of comparing your productwith competing products and your production and testing methods with those of yourcompetitors. The difference today is that it is now highly formalized and practisedwithout compunction. Once it is on the market any vehicle or component thereofcan be bought and tested by the manufacturer’s competitors, with a view to takingover and copying any features that are clearly in advance of the competitor’s ownproducts. There are test facilities built and run specifically for benchmarking.

This evidently increases the importance of patent cover, of preventing the transferof confidential information by disaffected employees and of maintaining confidential-ity during the development process; such concerns need to have preventative measuresbuilt into the specification of the facility rather than added as an afterthought.

Safety regulations and planning permits covering test cells

Feasibility not only concerns the technical and commercial viability, but also whetherone will be allowed to create the new or altered test laboratory; therefore, theresponsible person should consider discussion at an early stage with the followingagencies:

• local planning authority;• local petroleum officer and fire department;• local environmental officer;• building insurers;


• local electrical supply authority;• site utility providers.

Note the use of the word ‘local’. There are very few regulations specifically men-tioning engine test cells, much of the European legislation is generic and frequentlyhas unintended consequences for the automotive test industry. Most legislation isinterpreted locally and the nature of that interpretation will depend on the industrialexperience of the officials concerned, which can be highly variable. There is alwaysa danger that inexperienced officials will over-react to applications for engine testfacilities and impose unrealistic restraints on the design. It may be found useful tokeep in mind one basic rule that has had to be restated over many years:

An engine test cell, using liquid fuels, is a ‘zone 2’ hazard containment box. It is notpossible to make its interior inherently safe since the test engine worked to the extremesof its performance is not inherently safe; therefore the cell’s function is to contain andminimise the hazards and to inhibit human access when they are present. (See Chapter 4,Test cell and control room design: an overall view)

Most of the operational processes carried out within a typical automotive test cell aregenerally no more hazardous than those hazards experienced by garage mechanics,motorists or racing pit staff in real life. The major difference is that in the cell the run-ning engine is stationary in a space that is different from that for which it was designedand therefore humans may be able to gain close and potentially dangerous access to it.

It is more sensible to interlock the cell doors to prevent access to an engine runningabove ‘idle’ state, than to attempt to make the rotating elements ‘safe’ by the use ofclose fitting guarding that will inhibit operations and fall into operational disuse.

The authors of the high level operational specification need not concern themselveswith some of such details, but simply state that industrial best practice and compliancewith current legislation is required. The arbitrary imposition of existing operationalpractices on a new test facility should be avoided at the operational specificationstage until confirmed as appropriate, since they may restrict the inherent benefits ofthe technological developments available.

One of the restraints commonly imposed on the facility buildings concerns thenumber and nature of chimney stacks or ventilation ducts. This is often a causeof tension between the architect, planning authority and facility designers. Withsome ingenuity these essential items can be disguised, but the resulting designs willinevitably require more space than the basic vertical inlet and outlet ducts. Similarly,noise break-out via such ducting may be reduced to the background at the facilityborder but the space required for attenuation will complicate the plant room layout(see Chapter 3, Vibration and noise).

Note that the use of gaseous fuels will impose special restrictions on the design oftest facilities and, if included in the operational specification, the relevant authoritiesand specialist contractors must be involved from the planning stage. Modificationsmay include blast pressure relief panels in the cell structure and exhaust ducting,which need to be included from design inception.

6 Engine Testing

Specification for a control and data acquisition system

The choice of test automation supplier need not be part of the operation specificationbut, since it will form part of the functional specification, and since the choice oftest cell software may be the singularly most important technocommercial decisionin placing a contract for a modern test facility, it would seem sensible to considerthe factors that should be addressed in making that choice. The test cell automationsoftware lies at the core of the facility operation therefore its supplier will takean important role within the final system integration. The choice therefore is notsimply one of a software suite but of a key support role in the design and ongoingdevelopment of the new facility.

Laboratories where the systems are to be fully computerized should consider the

• local capability of each software/hardware supplier;• installed base of each possible supplier, relevant to the industrial sector;• level of operator training and support required for each of the short-listed systems;• compatibility of the control system with any intended, third party hardware;• modularity or upgradeability of both hardware and software;• requirements to use pre-existing data or to export data from the new facility to

existing databases;• ease of creating test sequences;• ease of channel calibration and configuration;• flexibility of data display and post-processing options.

A methodical approach allows for a ‘scoring matrix’ to be drawn up whereby com-peting systems may be objectively judged.

Anyone charged with producing specifications is well advised to carefully considerthe role of the test cell operators. Significant upgrades in test control and datahandling will totally change the working environment of the cell operator. There aremany cases of systems being imposed on users which never reach their full potentialbecause of inadequacy of training or inappropriate specification of the system.

Use of supplier’s specifications

It is all too easy for us to be influenced by headline speed and accuracy numbersin the specification sheets for computerized systems. The effective time constants ofmany engine test processes are not limited by the data handling rates of the computersystem, but rather of the physical process being measured and controlled. Thus thespeed at which a dynamometer can make a change in torque absorption is governedmore by the rate of magnetic flux generation in its coils, or the rate at which it canchange the mass of water in its internals, rather than the speed at which its controlalgorithm is being recalculated. The skill in using such information is to identify thenumbers that are relevant to task for which the item is required.

Faster is not necessarily better and it is often more expensive.


Functional specifications: some common difficulties

Building on the operational specification, which describes what the facility has todo, the functional specification describes how the facility is to perform its definedtasks and what it will contain. If the functional specification is to be used as thebasis for competitive tendering then it should avoid being unnecessarily prescrip-tive. Overprescriptive specifications, or those including sections that are technicallyincompetent, are problems to specialist contractors. The first type may prevent betteror more cost-effective solutions being quoted, while the later mean that a companywho, through lack of experience, claims compliance wins the contract, then inevitablyfails to meet the customer’s expectations.

Overprescription may range from ill matching of instrumentation to unrealisticallywide range of operation of subsystems.

A classic problem in facility specification concerns the range of engines that canbe tested in one test cell using common equipment and a single shaft system. Clearlythere is a great cost advantage for the whole production range of a manufacturer’sengines to be tested in one cell. However, the detailed design problems and sub-sequent maintenance implications that such a specification may impose can be fargreater than the cost of creating two or more cells that are optimized for narrowerranges of engines. Not only is this a problem inherent in the ‘turn-down’ ratio of fluidservices and instruments having to measure the performance of a range of enginesfrom say 500 to 60 kW, but the range of vibratory models produced may defy thecapability of any one shaft system to handle.

This issue of dealing with a range of vibratory models may require that cellsbe dedicated to particular types or that alternative shaft systems are provided forparticular engine types. Errors in this part of the specification and the subsequentdesign strategy are often expensive to resolve after commissioning. Not even themost demanding customer can break the laws of physics with impunity.

Before and during the specification and planning stage of any test facility, allparticipating parties should keep in mind the vital question: By what cost- and time-effective means do we prove that this complex facility meets the requirement andspecification of the user? It is never too early to consider the form and contentof acceptance tests, since from them the designer can infer much of the detailedfunctional specification. Failure to incorporate these into contract specifications fromthe start leads to delays and disputes at the end.

Interpretation of specifications

Employment of contractors with the relevant industrial experience is the best safe-guard against overblown contingencies or significant omissions in quotations arisingfrom user-generated specifications.

8 Engine Testing

Provided with a well written operational and functional specification any compe-tent subcontractor experienced in the engine or vehicle test industry should be able toprovide a detailed specification and quote for their module or service within the totalproject. Subcontractors who do not have experience in the industry will not be ableto appreciate the special, sometimes subtle, requirements imposed upon their designsby the transient conditions, operational practices and possible system interactionsinherent in the industry.

In the absence of a full appreciation of the project based on previous experiencethey will search the specification for ‘hooks’ on which to hang their standard productsor designs, and quote accordingly. This is particularly true of air or fluid conditioningplant where the bare parameters of temperature range and heat load can lead theinexperienced to equate test cell conditioning with that of a chilled warehouse. Anescorted visit to an existing test facility should be the absolute minimum experiencefor subcontractors quoting for systems such as chilled water, electrical installationand HVAC.

General project organization

In all but the smallest test facility projects, there will be three generic types ofcontractor with whom the customer’s project manager has to deal. They are

• civil contractor;• building services contractors;• test instrumentation contractor.

How the customer decides to deal with these three industrial groups and integrate theirwork will depend on the availability of in-house skills and the skills and experienceof any preferred contractors.

The normal variations in project organization, in ascending order of customerinvolvement in the process, are

• a consortium working within a design and build or ‘turnkey’∗ contract based onthe customer’s operational specification and working to the detailed functionalspecification and fixed price produced by the consortium;

• guaranteed maximum price (GMP) contracts where a complex project manage-ment system, having a ‘open’ cost accounting system, is set up with the mutualintent to keep the project within a mutually agreed maximum value. This requiresjoint project team cohesion of a high order;

∗ The term turnkey is now widely misused. The original turnkey contract was one carriedout to an agreed specification by a contractor taking total responsibility for the site and allassociated works with virtually no involvement by the end user until the keys were handedover so that acceptance tests can be performed.


• a customer-appointed main contractor employing supplier chain and working tocustomer’s functional specification;

• a customer appointed civil contractor followed by services and system inte-grator contractor each appointing specialist subcontractors, working with cus-tomer’s functional specification and under the customer’s project managementand budgetary control;

• a customer controlled series of subcontract chains working to the cus-tomer’s detailed functional specification, project engineering, site and projectmanagement.

Whichever model is chosen the two vital roles of project manager and design authority(systems integrator) have to be clear to all and provided with the financial andcontractual authority to carry out their allotted roles. It should be noted that inthe UK, all but the smallest contracts involving construction or modification oftest facilities will fall under the control of a specific section of health and safetylegislation known as Construction Design and Management Regulations 1994 (CDMRegulations) which require nomination of these and other project roles.

Project roles and management

The key role of the client, or user, is to invest great care and effort into the creation ofa good operational and functional specification. Once permission to proceed has beengiven, based on this specification, and the main contractor has been appointed, theday to day role of the client user group should, ideally, reduce to that of attendanceat review meetings and being ‘on-call’.

Nothing is more guaranteed to cause project delays and cost escalation than ill-considered or informal changes of detail by the client’s representatives. Whateverthe project model, the project management system should have a formal systemof notification of change and an empowered group within both the customer’s andcontractor’s organization to deal with such changes quickly. The type of form shownin Fig. 1.1 allows individual requests for project change to be recorded and theimplications of the change to be discussed and quantified. Change can have negativeor positive effect on project costs and can be requested by both the client and thecontractor; with the right working relationship in a joint project team, change notescan be the mechanism by which cost- or time-saving alterations can be raised duringthe course of work.

All projects have to operate within the three restraints of time, cost and qual-ity (content) (Fig. 1.2). The relative importance of these three criteria has to beunderstood by the client and project manager. The model is different for eachclient and each project and however much a client may protest that all three cri-teria have equal weighting and are fixed; if change is introduced, one has to be avariable.

10 Engine Testing

Customer: Variation No:

Project name: Project No:

Details of proposed change: (include any reference to supporting documentation)

Requirement (tick and initial as required)

Design authority required design change Customer instruction: Customer request: Contractor request: Urgent quotation required: Customer agreed to proceed at risk: Work to cease until variation agreed: Requested to review scope of supply: Other (specify):

Name Email / Phone No.

Authorized

Customer representative

Authorized

Actions:Sales quote submitted: (date and initial)Authorized for action: (date and initial)Implemented: (date and initial)

Name Email / Phone No.

Contractorrepresentative

Figure 1.1 A sample contract variation record sheet

If one of the points of the triangle is moved there is a consequential change inone or both of the others and that the later in the program the change is required, thegreater the consequential effect.

One oft-repeated error, which is forced by time pressure on the overall programme,is to deliver instrumentation and other equipment into a facility building beforeinternal environmental conditions are suitable. It is always better to deliver such plantlate and into a suitable environment, then make up time by increasing installationman hours, than it is to have incompatible trades working in the same building spaceand suffering the almost inevitable damage to the test equipment.


Costof project

Time(programme)

Contentquality

3project

variables

Figure 1.2 Project constraints: in real life, if two are fixed the third will bevariable

Key project management tools

The techniques of project management are outside the remit of this book but someimportant tools and skills suitable for any project manager in charge of an automotivetest facility are suggested.

Communications and responsibility matrix

Any multicontractor and multidisciplinary project creates a complex network of com-munications. These networks between suppliers, contractors and personnel withinthe customer’s organization may pre-exist or be created during the course of theproject; the danger is that informal communications may give rise to unauthorizedvariations in project content or timing. Good project management is only possi-ble with a disciplined communication system and this should be designed intoand maintained during the project. The arrival of email as the standard commu-nication method has increased the need for communication discipline and intro-duced the need, within project teams, of creating standardized, computer-based filingsystems.

If competent staff are available to create and maintain a project specific intranetwebsite then the project manager has a good means of maintaining control overformal communications. Such a network can give access permission, such as ‘readonly’, ‘submit’ and ‘modify’, as appropriate to individuals’ roles to all of thegroups and nominated staff having any commercial or technical interest in theproject.

The creation of a responsibility matrix is most useful when it covers the importantminutiae of project work, that is, not only who supplies a given module but whoinsures, delivers, off-loads, connects and commissions the module.

12 Engine Testing

Use of ‘master drawing’ in project control

The use of a facility layout or schematic drawing graphically showing facility modulesthat can be used by all tendering contractors and continually updated by the maincontractor or design authority can be a vital tool in any multidisciplinary projectwhere there may be little detailed appreciation between specialized contractors forthe spatial requirements of each other.

Constant, vigilant site management is required during the final building ‘fit out’phase of an automotive test facility if clashes over space allocation are to be avoided,but good preparation and contractor briefing can reduce the inherent problem. Ifthe systems integrator or main contractor takes ownership of project floor layoutplans and these plans are used at every subcontractor meeting, kept up to dateto record the layout of all services and major modules, then most of the spaceutilization, service route and building penetration problems will be resolved beforework commences. Where possible and appropriate, contractors method statementsshould use the common ‘table top’ project plans to show the area of their owninstallation in relation to the building and installations of others.

Project timing chart

Most staff involved with a project will recognize a classic Gantt chart; not all willunderstand their role or the interactions of their tasks within that plan. It is the taskof the project manager to ensure that each contractor and all key personnel workwithin the project plan structure. This is not served by sending an electronic versionof a large and complex Gantt chart, but by early contract briefing and preinstallationprogress meetings.

There are some key events in every project that are absolutely time critical andthese have to be given special attention by both client and project manager. Consider,for example, the arrival of a chassis dynamometer and the site implications:

• One or more large trucks will have to arrive on the client’s site, in the correctorder, and require suitable site access for manoeuvring.

• The chassis dynamometer will require a large crane to off-load. The crane’sarrival and site positioning will have to be coordinated within an hour of thetrucks’ arrival.

• Access into the chassis dynamometer pit area will have to be kept clear for specialheavy handling equipment until the unit is installed, the access thereafter will beclosed up by deliberately delayed building work.

• Other contractors will have to be kept out of the effected work and access areas,as will client’s and contractor’s vehicles and equipment.

Preparation for such an event takes detailed planning, good communications andauthoritative management. The non- or late arrival of one of the key players because‘they did not understand the importance’ clearly causes acute problems in the


example above, but the same ignorance of programmed roles causes delays andoverspends that are less obvious throughout any project where detailed planning andcommunications are left to take care of themselves.

A note on documentation

Test cells and control room electrical systems are, in the nature of things, sub-ject to detailed modification during the build and commissioning process. Thedocumentation, representing the ‘as-commissioned’ state of the facility, must be ofa high standard and easily accessible to maintenance staff and contractors. The formand due delivery of documentation should be specified within the functional specifi-cation and form part of the acceptance criteria. Subsequent responsibility for keepingrecords and schematics up to date within the operator’s organization must be clearlydefined and controlled.

Summary

Like all complex industrial projects, the creation, or significant modification, of anengine test laboratory should start with the creation of an operational specificationinvolving all those departments and individuals having a legitimate interest. Thespecification of the engine testing tasks and definition of suitable acceptance testsare essential prerequisites of such a project.

The roles of system integrator, design authority and project manager must be welldefined and empowered. Specifically, ensure that the design authority for systemsintegration is explicitly given to a party technically competent and contractuallyempowered to carry it out.

There are at least three levels of specification that should be considered and it isessential to invest time in their preparation if a successful project is to be achieved.Paper is cheaper to change than concrete.

The project management techniques required are those of any multidisciplinarylaboratory construction but require knowledge of the core testing process so that themany subtasks are integrated appropriately.

The statement made early in this chapter, ‘Without a clear and unambiguous spec-ification no complex project should be allowed to proceed’, seems self-evident; yetmany companies, within and outside our industry, continue either to allocate the taskinappropriately or underestimate its importance and subject it to post-order change.The consequence is that project times are extended by an iterative quotation periodand from the point at which the users realize that their (unstated or misunderstood)expectations are not being met, usually during commissioning.

2 The test cell as a thermodynamicsystem

The energy of the world is constant; the entropy strives towards a maximum.1

Rudolph Clausius (1822–1888)

Introduction

The closed engine test cell system makes a suitable case for students to study anexample of the flow of heat and change in entropy. In almost all engine test cells thevast majority of the energy comes into the system as highly concentrated ‘chemicalenergy’ entering the cell via the smallest penetration in the cell wall, the fuel line.It leaves the cell as lower grade heat energy via the largest penetrations: the ventilationduct, engine exhaust pipe and the cooling water pipes. In the case of cells fittedwith electrically regenerative dynamometers, almost one-third of the energy suppliedby fuel will leave the cell as electrical energy able to slow down the electricalenergy supply meter. Many problems are experienced in test cells worldwide whenthe thermodynamics of the cell have not been correctly catered for in the design ofcooling systems. The most common problem is high air temperature within the testcell, either generally or in critical areas. The practical effects of such problems willbe covered in detail in Chapter 5, Ventilation and air conditioning, but it is vital forthe cell designer to have a general appreciation of the contribution of the variousheat sources and the strategies for their control.

In the development of the theory of thermodynamics much use is made of theconcept of the open system. This is a powerful tool and can be very helpful inconsidering the total behaviour of a test cell. It is linked to the idea of the controlvolume, a space enclosing the system and surrounded by an imaginary surface, thecontrol surface (Fig. 2.1).

The great advantage of this concept is that once one has identified all the massand energy flows into and out of the system it is not necessary to know exactly whatis going on inside the system in order to draw up a ‘balance sheet’ of inflows andoutflows.

The test cell as a thermodynamic system 15

Control surface

Control volume

In

Fuel

Air

Water

Out

Machine

Exhaust

Power

Water

Figure 2.1 An open thermodynamic system

The various inflows and outflows to and from a test cell are as follows:

In∗ Out

FuelVentilation air (some may be used bythe engine as combustion air)

Ventilation air

Combustion air (treated) Exhaust (includes air used by engine)Engine cooling water

Charge air (when separately supplied) Dynamometer cooling waterCooling water Electricity from dynamometerElectricity for services Losses through walls and ceiling

Balance sheets may be drawn up for fuel, air, water and electricity, but by farthe most important is the energy balance, since every one of these quantities hasassociated with it a certain quantity of energy. The same concept may be applied tothe engine within the cell. This may be pictured as surrounded by its own controlsurface, through which the following flows take place:

In Out

Fuel PowerAir used by the engine ExhaustCooling water Cooling waterCooling air Cooling air

Convection and radiation

∗ Compressed air may be a further energy input; however, usage is generally intermittent andunlikely to make a significant contribution.

16 Engine Testing

Measurement of thermal losses from the engine is dealt with in Chapter 13, Thermalefficiency, measurement of heat and mechanical losses, where the value of the methodin the analysis of engine performance is made clear.

The energy balance of the engine

Table 2.1 shows a possible energy balance sheet for a cell in which a gasoline engineis developing a steady power output of 100 kW. Note that where fluids (air, water,exhaust) are concerned, the energy content is referred to an arbitrary zero, the choiceof which is unimportant: we are only interested in the difference between the variousenergy flows into and out of the cell.

Given sufficient detailed information on a fixed engine/cell system it is possibleto carry out a very detailed energy balance calculation (see Chapter 5, Ventilationand air conditioning, for a more detailed treatment). Alternatively there are somecommonly used ‘rule of thumb’ calculations available to the cell designer; the mostcommon of these relates to the energy balance of the engine which is known as the‘30–30–30–10 rule’. This refers to the energy balance shown in Table 2.2.

The key lesson to be learnt by the non-specialist reader is that any engine testcell has to be designed to deal with energy flows that are at least three times greaterthan the ‘headline’ engine rating. To many this will sound obvious but a commonfixation on engine power and a casual familiarity with, but lack of appreciation of, theenergy density of petroleum fuels has misled many people in the past to significantly

Table 2.1 Simplified energy flows for a test cell fitted with a hydraulicdynamometer and 100 kW gasoline engine

Energy balance

In Out

Fuel 300 kW Exhaust gas 60 kWVentilating fan power 5 kW Engine cooling water 90 kW

Dynamometer cooling water 95 kWVentilation air 70 kW

Electricity for cell services 25 kW Heat loss, walls and ceiling 15 kW

330 kW 330 kW

The energy balance for the engine, see Chapter 11, is as follows:

In Out

Fuel 300 kW Power 100 kWExhaust gas 90 kWEngine cooling water 90 kWConvection and radiation 20 kW

300 kW 300 kW


Table 2.2 Example of the 30–30–30–10 rule

In via Out via

Fuel 300 kW Dynamometer 30% (90+ kW)Exhaust system 30% (90 kW)Engine fluids 30% (90 kW)Convection and radiation 10% (30 kW)

under-rate cell cooling systems. Like any rule of thumb this is crude but does providea starting point for the calculation of a full energy balance and a datum from whichwe can evaluate significant differences in balance caused by the engine itself and itsmounting within the cell.

Firstly, there are differences inherent in the engine design. Diesels will tend totransfer less energy into the cell than petrol engines of equal physical size. Forexample, testers of rebuilt bus engines often notice that different models of dieselswith the same nominal power output will show quite different distribution of heatinto the test cell air and cooling water.

Secondly, there are differences in engine rigging in the cell which will vary thetemperature and surface area of engine ancillaries, such as exhaust pipes. Finally,there is the amount and type of equipment within the test cell, all of which makes acontribution to the convection and radiation heat load to be handled by the ventilationsystem.

Specialist designers have developed their own versions of a software model, basedboth on empirical data and theoretical calculation, all of which is used within thisbook, and which produces the type of energy balance shown in Fig. 2.2. Such toolsare useful but cannot be used uncritically as the final basis of design, particularlywhen a range of engines are to be tested or the design has to cover two or morecells, then the energy diversity factor has to be considered.

Diversity factor and the final specification of a facilityenergy balance

To design a multicell test laboratory able to control and dissipate the maximumtheoretical power of all its prime movers on the hottest day of the year will lead toan oversized system and possibly poor temperature control at low heat outputs. Theamount by which the thermal rating of a facility is reduced from that theoretical maxi-mum is the diversity factor. In Germany it is called the Gleichzeitigkeits Faktor and iscalculated from zero heat output upwards, rather than 100 per cent heat output down-wards, but the results should be the same, providing the same assumptions are made.

The diversity factor often lies between 60 and 85 per cent of maximum rating, butindividual systems will vary from endurance beds with high rating down to anechoicbeds with very low rating.

Power output

CFM

CFM

LightsFuel

Cooling water to room0.9%

Process water forintercooler

Process water foroil cooling

Process water fordyno coolingCombustion air

8.0%

38

Radiation energy

from engine

Ventilation air

Exhaust gas

Exhaust radiation energy

Heat from dyno

cellAC dyno

m3/s

12153

91

1.2

28

6.0% 3.2%

15

135

3.8

Electrical power

Electrical service

5

19.5%

12153

5.7

40

919

0.43

150 kW

kW

kW kW

kW

kW

kW

kW

kW

kWkW

kW4

kW kW

kW

m

Power output

Instrumentation

m wide

kW

°C

°C

°C

CFM

m3/s CFM

m3/s

°C

m3/s

°C

kW

kW

kg/kW.h

I/h

469

32.0%

0.262

47.3

0.0%

0.0

33

29.1%

136

55

158

96

5.7

26

499 21

5

6

150

0 0

0.0% 0.0%4.5%0.24

25

Energy input

Efficiency

Fuel consumption

Fuel cooling

Total processwater heat load

Process water forengine cooling

Total heat loadinto cell

(inc. lighting andbuilding gains)

Engine Dyno

Figure 2.2 Output diagram from test cell thermal analysis software


0

0.2

0.4

0.6

0.8

1

1 2 3 4 5 6

Number of test cells

Div

ers

ity f

acto

r

Figure 2.3 Diversity factor of thermal rating of facility services, plotted againstnumber of test cells, based on empirical data from typical automotive test facilities

In calculating, or more correctly estimating, the diversity factor it is essential thatthe creators of the operational and functional specifications use realistic values ofactual engine powers, rather than extrapolations based on possible future develop-ments. A key consideration is the number of cells included within the system. Itis clear that one cell may at some time run at its maximum rating, but it may beconsidered less likely that four cells will all run at maximum rating at the sametime: the possible effect of this is shown in Fig. 2.3. There is a degree of braveryand confidence, based on relevant experience, required to significantly reduce thetheoretical maximum to a contractual specification, but very significant savings infirst and running costs may be possible if it is done correctly.

Once a realistic maximum power rating for the facility has been calculated, thefacility design team can use information concerning the operating regime, plannedtest sequences, past records, engine type variation, etc., to draw up diversity factorsfor heat energy balance and electrical power requirements. Future proofing may bebetter designed into the facility by incremental addition of plant rather than oversizingat the beginning.

Common or individual services in multicell laboratories?

When considering the thermal loads and the diversity factor of a facility containingseveral test cells, it is sensible to consider the strategy to be adopted in the design ofthe various services. The choice has to be based on the operation requirements ratherthan just the economies of purchasing and running the service modules. Services,such as cooling water (raw water), are always common and treated via a centralcooling tower system. Services, such as cell ventilation and engine exhaust gasextraction, may either serve individual cells or are shared. In these cases sharingmay show cost savings and simplify the building design by reducing penetrations;

20 Engine Testing

however, it is prudent to build in some standby or redundancy to prevent total facilityshut-down in the event of, for example, a fan failure.

A problem that must be avoided in the design of common services is ‘cross talk’between cells where the action in one cell, or other industrial plant, disturbs thecontrol achieved in another. This is a particular danger when a service, for examplechilled water, has to serve a wide range of thermal loads. In this case a central plantmay be designed to circulate glycol/water mix at 6�C through two or more cellswherein the coolant is used by devices ranging from large intercoolers to small fuelconditioners; any sudden increase in demand may significantly increase the systemreturn temperature and cause an unacceptable disturbance in the control temperatures.In such systems there needs to be individual control loops per instrument or avery high thermal inertia gained through the installation of a sufficiently large coldbuffer tank.

Summary

The energy balance approach outlined in this chapter will be found helpful inanalysing the performance of an engine and in the design of test cell services (Chap-ters 5 and 6). It is recommended that at an early stage in the design of a new test cell,diagrams such as Fig. 2.2 should be drawn up and labelled with flow and energyquantities appropriate to the capacity of the engines to be tested.

The large quantities of ventilation air, cooling water, electricity and heat that areinvolved will often come as a surprise. Early recognition can help to avoid expensivewasted design work by ensuring that

• the general proportions of cell and services do not depart too far from acceptedpractice (any large departure is a warning sign);

• the cell is made large enough to cope with the energy flows involved;• sufficient space is allowed for such features as water supply pipes and drains, air

inlet grilles, collecting hoods and exhaust systems. Note that space is not onlyrequired within the test cell but also in any service spaces above or below thecell and the penetrations within the building envelope.

Reference

1. Eastop, T.D. and McConkey, A. (1993) Applied Thermodynamics for Engineering Tech-

nologists, Longmans, London.

Further reading

Heywood, J.B. (1988) Internal Combustion Engine Fundamentals, McGraw-Hill,Maidenhead.

3 Vibration and noise

Introduction

Vibration is considered in this chapter with particular reference to the design and oper-ation of engine test facilities, engine mountings and the isolation of engine-induceddisturbances. Torsional vibration is covered as a separate subject in Chapter 9, Cou-pling the engine to the dynamometer.

The theory of noise generation and control is briefly considered and a brief accountgiven of the particular problems involved in the design of anechoic cells.

Vibration and noise

Almost always the engine itself is the only significant source of vibration and noisein the engine test cell.1−5 Secondary sources such as the ventilation system, pumpsand circulation systems or the dynamometer are usually swamped by the effects ofthe engine.

There are several aspects to this problem:

• The engine must be mounted in such a way that neither it nor connections to itcan be damaged by excessive movement or excessive constraint.

• Transmission of engine-induced vibration to the cell structure or to other buildingsmust be controlled.

• Excessive noise levels in the cell should be avoided or contained as far as possibleand the design of alarm signals should take in-cell noise levels into account.

Fundamentals: sources of vibration

Since the vast majority of engines likely to be encountered are single- or multi-cylinder in-line vertical engines, we shall concentrate on this configuration.

An engine may be regarded as having six degrees of freedom of vibration aboutorthogonal axes through its centre of gravity: linear vibrations along each axis androtations about each axis (see Fig. 3.1).

22 Engine Testing

Z

Z

X

X

Y

Y

Figure 3.1 Internal combustion engine: principle axes and degrees of freedom

In practice, only three of these modes are usually of importance:

• vertical oscillations on the X axis due to unbalanced vertical forces;• rotation about the Y axis due to cyclic variations in torque;• rotation about the Z axis due to unbalanced vertical forces in different transverse

planes.

Torque variations will be considered later. In general, the rotating masses are carefullybalanced but periodic forces due to the reciprocating masses cannot be avoided. Thecrank, connecting rod and piston assembly shown in Fig. 3.2 is subject to a periodicforce in the line of action of the piston given approximately by:

f =mp�2cr cos�+

mp�2cr cos2�

nwhere n= l/r (1)

m

f

I

r

ωc

θ

Figure 3.2 Connecting rod crank mechanism: unbalanced forces

Vibration and noise 23

Here mp represents the sum of the mass of the piston plus, by convention, one-thirdof the mass of the connecting rod (the remaining two-thirds is usually regarded asbeing concentrated at the crankpin centre).

The first term of eq. (1) represents the first-order inertia force. It is equivalentto the component of centrifugal force on the line of action generated by a mass mp

concentrated at the crankpin and rotating at engine speed. The second term arisesfrom the obliquity of the connecting rod and is equivalent to the component of forcein the line of action generated by a mass m/4n at the crankpin radius, but rotating attwice engine speed.

Inertia forces of higher order (3×, 4×, etc., crankshaft speed) are also generatedbut may usually be ignored.

It is possible to balance any desired proportion of the first-order inertia forceby balance weights on the crankshaft, but these then give rise to an equivalentreciprocating force on the Z axis, which may be even more objectionable.

Inertia forces may be represented by vectors rotating at crankshaft speed and twicecrankshaft speed. Table 3.1 shows the first- and second-order vectors for engineshaving from one to six cylinders.

Table 3.1 First- and second-order forces, multicylinder engines

Firstorderforces

1

1 1 1 1 1 1

1 1

1

1

1 1 1 1 1

1

1

1

1

1

1

1

1

1

1

1

1

22

22 2

2

2

2

2

2

2

2

2

2

2

33 3 3

3

3

3

5

5

5

5

5

6

6

6

6

5

5

5

33

3

3

3

3

3

2 2 2 2 2

3 3 3 3

4

4 4

5

56

4

4

4

4

4

4

4

4

4

4

4

4

2

2

2

Secondorderforces

Firstordercouples

Secondordercouples

24 Engine Testing

Note the following features:

• In a single cylinder engine, both first- and second-order forces are unbalanced.• For larger numbers of cylinders, first-order forces are balanced.• For two and four cylinder engines, the second-order forces are unbalanced and

additive.

This last feature is an undesirable characteristic of a four cylinder engine and in somecases has been eliminated by counter-rotating weights driven at twice crankshaftspeed.

The other consequence of reciprocating unbalance is the generation of rockingcouples about the transverse or Z axis and these are also shown in Fig. 3.1.

• There are no couples in a single cylinder engine.• In a two cylinder engine, there is a first-order couple.• In a three cylinder engine, there are first- and second-order couples.• Four and six cylinder engines are fully balanced.• In a five cylinder engine, there is a small first-order and a larger second-order

couple.

Six cylinder engines, which are well known for smooth running, are balanced in allmodes.

Variations in engine turning moment are discussed in Chapter 9, coupling theengine to the dynamometer. These variations give rise to equal and opposite reactionson the engine, which tend to cause rotation of the whole engine about the crankshaftaxis. The order of these disturbances, i.e. the ratio of the frequency of the disturbanceto the engine speed, is a function of the engine cycle and the number of cylinders.For a four-stroke engine, the lowest order is equal to half the number of cylinders:in a single cylinder there is a disturbing couple at half engine speed while in a sixcylinder engine the lowest disturbing frequency is at three times engine speed. In atwo-stroke engine, the lowest order is equal to the number of cylinders.

The design of engine mountings and test bed foundations

The main problem in engine mounting design is that of ensuring that the motionsof the engine and the forces transmitted to the surroundings as a result of theunavoidable forces and couples briefly described above are kept to manageablelevels. In the case of vehicle engines it is sometimes the practice to make use of thesame flexible mounts and the same location points as in the vehicle; this does not,however, guarantee a satisfactory solution. In the vehicle, the mountings are carriedon a comparatively light structure, while in the test cell they may be attached to amassive pallet or even to a seismic block. Also in the test cell the engine may befitted with additional equipment and various service connections. All of these factorsalter the dynamics of the system when compared with the situation of the engine in


service and can give rise to fatigue failures of both the engine support brackets andthose of auxiliary devices, such as the alternator.

Truck diesel engines usually present less of a problem than small automotiveengines, as they generally have fairly massive and well-spaced supports at the fly-wheel end. Stationary engines will in most cases be carried on four or more flexiblemountings in a single plane below the engine and the design of a suitable system isa comparatively simple matter.

We shall consider the simplest case, an engine of mass m kg carried on undampedmountings of combined stiffness k N/m (Fig. 3.3). The differential equation definingthe motion of the mass equates the force exerted by the mounting springs with theacceleration of the mass:

md2x

dt2+kx = 0 (2)

a solution is

x = cons tan t× sin

√

k

m· t

k

m= �2

0 natural frequency= �0 =�0

2�=

1

2�

√

k

m(3)

the static deflection under the force of gravity=mg/kwhich leads to a very convenientexpression for the natural frequency of vibration:

�0 =1

2�

√

g

static deflection(4a)

C of G

Figure 3.3 Engine carried on four flexible mountings

26 Engine Testing

or, if static deflection is in millimetres:

�0 =15�76

√static deflection

(4b)

This relationship is plotted in Fig. 3.4Next, consider the case where the mass m is subjected to an exciting force of

amplitude f and frequency �/2�. The equation of motion now reads:

md2x

dt2+kx = f sin�t

the solution includes a transient element; for the steady state condition amplitude ofoscillation is given by:

x =f/k

�1−�2/�20�

(5)

here f / k is the static deflection of the mountings under an applied load f. Thisexpression is plotted in Fig. 3.5 in terms of the amplitude ratio x divided by staticdeflection. It has the well-known feature that the amplitude becomes theoreticallyinfinite at resonance, �= �0.

If the mountings combine springs with an element of viscous damping, the equa-tion of motion becomes:

md2x

dt2+ c

dx

dt+kx = f sin�t

40

30

20

10

54

3

2

10.2

Natu

ral fr

equency (

Hz)

0.5 1 2 5

Static deflection (mm)

10 20 50 100 300

Figure 3.4 Relationship between static deflection and natural frequency


6

4

2

00

Am

plit

ud

e r

atio

1 2

Frequency ratioω0

ω

Figure 3.5 Relationship between frequency and amplitude ratio (transmissibil-ity) undamped vibration

where c is a damping coefficient. The steady state solution is:

x =f/k

√

(

1−�2

�20

)2

+�2c2

mk�20

sin��t−A� (6a)

If we define a dimensionless damping ratio:

C2 =c2

4mk

this equation may be written:

x =f/k

√

(

1−�2

�20

)2

+4C2�2

�20

sin��t−A� (6b)

(if C= 1 we have the condition of critical damping when, if the mass is displacedand released, it will return eventually to its original position without overshoot).

The amplitude of the oscillation is given by the first part of this expression:

amplitude=f/k

√

(

1−�2

�20

)2

+4C2�2

�20

28 Engine Testing

100

0.05

0.1

0.1

0

0.2

0.2

0.5

0.5

0.05

5

2

1

0.8

0.6

0.4

0.3

0.2

0.10.08

0.06

0.04

0.03

0.02

00.1 0.2

Tra

nsm

issib

ility

0.5 1 2 3 4 5 7 10

Frequency ratioω0

ω

C

Figure 3.6 Relationship between transmissibility (amplitude ratio) and fre-quency, damped oscillations for different values of damping ratio C (logarithmicplot)

This is plotted in Fig. 3.6, together with the curve for the undamped condition,Fig. 3.5, and various values of C are shown. The phase angle A is a measure of theangle by which the motion of the mass lags or leads the exciting force. It is givenby the expression:

A= tan−1 2C�00

�−

�

�0

(7)

At very low frequencies, A is zero and the mass moves in phase with the exciting force.With increasing frequency the motion of the mass lags by an increasing angle, reach-ing 90� at resonance. At high frequencies the sign of A changes and the mass leadsthe exciting force by an increasing angle, approaching 180� at high ratios of� to�0.

Natural rubber flexible mountings have an element of internal (hysteresis) dampingwhich corresponds approximately to a degree of viscous damping corresponding toC= 0.05.


The essential role of damping will be clear from Fig. 3.6: it limits the potentiallydamaging amplitude of vibration at resonance. The ordinate in Fig. 3.6 is oftendescribed as the transmissibility of the mounting system: it is a measure of the extentto which the disturbing force f is reduced by the action of the flexible mounts. It isconsidered good practice to design the system so that the minimum speed at whichthe machine is to run is not less than three times the natural frequency, correspondingto a transmissibility of about 0.15. It should be noticed that once the frequencyratio exceeds about 2 the presence of damping actually has an adverse effect on theisolation of disturbing forces.

Practical considerations in the design of engine and test bedmountings

In the above simple treatment we have only considered oscillations in the verticaldirection. In practice, as has already been pointed out, an engine carried on flexiblemountings has six degrees of freedom (Fig. 3.1). While in many cases a simpleanalysis of vibrations in the vertical direction will give a satisfactory result, under testcell conditions a more complete computer analysis of the various modes of vibrationand the coupling between them may be advisable. This is particularly the case withtall engines with mounting points at a low level, when cyclic variations in torquemay induce transverse rolling of the engine.

Reference 6 lists the design factors to be considered in planning a system for theisolation and control of vibration and transmitted noise:

• specification of force isolation

– as attenuation, dB– as transmissibility– as efficiency– as noise level in adjacent rooms

• natural frequency range to achieve the level of isolation required• load distribution of the machine

– is it equal on each mounting?– is the centre of gravity low enough for stability?– exposure to forces arising from connecting services, exhaust system, etc.

• vibration amplitudes – low frequency

– normal operation– fault conditions– starting and stopping– is a seismic block or sub-base needed?

30 Engine Testing

• higher-frequency structure-borne noise (100Hz+)

– is there a specification?– details of building structure– sufficient data on engine and associated plant

• transient forces

– shocks, earthquakes, machine failures

• environment

– temperature– humidity– fuel and oil spills.

Detailed design of engine mountings for test bed installation is a highly specializedmatter, see Ker-Wilson1 for guidance on standard practice. In general, the aim is toavoid ‘coupled’ vibrations, e.g. the generation of pitching forces due to unbalancedforces in the vertical direction, or the generation of rolling moments due to the torquereaction forces exerted by the engine. These can give rise to resonances at muchhigher frequencies than the simple frequency of vertical oscillation calculated in thefollowing section and to consequent trouble, particularly with the engine-to-brakeconnecting shaft.

Massive foundations and supported bedplates

The analysis and prediction of the extent of transmitted vibration to the surroundingsis a highly specialized field. The theory is dealt with in Ref. 1, the starting point beingthe observation that a heavy block embedded in the earth has a natural frequencyof vibration that generally lies within the range 1000 to 2000 c.p.m. There is thusa possibility of vibration being transmitted to the surroundings if exciting forces,generally associated with the reciprocating masses in the engine, lie within thisfrequency range. An example would be a four cylinder four-stroke engine runningat 750 rev/min. We see from Table 3.1 that such an engine generates substantialsecond-order forces at twice engine speed or 1500 c.p.m. Figure 3.7, redrawn fromRef. 1, gives an indication of acceptable levels of transmitted vibration from thepoint of view of physical comfort.

Figure 3.8 is a sketch of a typical seismic block. Reinforced concrete weighsroughly 2500 kg/m3 and this block would weigh about 4500 kg. Note that the sur-rounding tread plates must be isolated from the block, also that it is essential toelectrically earth (ground) the mounting rails. The block is shown carried on fourcombined steel spring and rubber isolators, each having a stiffness of 100 kg/mm(Fig. 3.9). From eq. 2a, the natural frequency of vertical oscillation of the bareblock would be 4.70Hz or 282 c.p.m., so the block would be a suitable base for anengine running at about 900 rev/min or faster. If the engine weight were, say, 500 kg


7

2

1

0.7

0.2

0.1

0.07

0.02

0.01

0.007

0.002

100 200 500

Rough

Normal

Smooth

Perceptible

Imperceptible

Frequency c.p.m.

Am

plit

ude m

m

1000 2000 5000

Figure 3.7 Perception of vibration

+

–

12 m

0.8 m

Figure 3.8 Spring-mounted seismic block

32 Engine Testing

Resilient pads

Resilient pad

Assembly bolt and snubberSlot to inspectsnubber position

Base casting

Top casting

Rubber spring

Helical steel spring

Figure 3.9 Combined spring and rubber flexible mount

the natural frequency of block and engine would be reduced to 4.46Hz, a negligiblechange. An ideal design target for the natural frequency is considered to be 3Hz.

Heavy concrete foundations (seismic blocks) carried on a flexible membrane areexpensive to construct, calling for deep excavations, complex shuttering and elaboratearrangements, such as tee-slotted bases, for bolting down the engines. With the widerange of different types of flexible mounting now available, it is questionable whether,except in special circumstances, such as a requirement to install test facilities in closeproximity to offices, their use is economically justified. The trough surrounding theblock may be of incidental use for installing services, if the gap is small then thereshould be means of draining out contaminated fluid spills.

It is now common practice for automotive engines to be rigged on vehicle typeengine mounts, then on trolley systems, which are themselves mounted on isolationfeet therefore less engine vibration is transmitted to the building floor. In thesecases a more modern alternative to the deep seismic block is shown in Fig. 3.10aand is sometimes used where the soil conditions are suitable. Here the test bedsits on a thickened and isolated section of the floor cast in situ on the compactednative ground. The gap between the floor and block is almost filled with expandedpolystyrene boards and sealed at floor level with a flexible, fuel resistant, sealant. Adamp-proof membrane should be inserted under both floor and pit or block.

Where the subsoil is not suitable for the arrangement shown in Fig. 3.10a thena pit is required, cast to support a concrete block that sits on a mat or pads of amaterial such as cork/nitrile rubber composite which is resistant to fluid contamination(Fig. 3.10b), alternatively a cast iron bedplate supported by air springs, as shown inFigure 3.11, may be installed.


(a)

Consolidated subsoil

Cell floor

Seismic block

Isolation gap filled with fuel resistant sealant

Cell floorSeismic block

placed on

resilient mat

Concrete pit

(b)

Figure 3.10 (a) Isolated foundation block for test stand set on to firm subsoil;(b) seismic block on to resilient matting in a shallow pit

4

2

1

53

Figure 3.11 Seismic block or cast bedplate (not shown) (4) mounted on airsprings (3) within a shallow concrete pit (2) on consolidated subsoil (1). Whenthe air supply is off the self-levelling air springs allow the block to settle onsupport blocks (5) the rise and fall typically being 4 to 6mm. Maintenanceaccess to the springs is not shown

34 Engine Testing

The use of a cast iron or steel fabricated bedplate mounted on springs is widespreadand air-spring support system of four or more units is connected so that, aftercommissioning, the system is self-levelling so that the base plate remains in a constantload-independent vertical position and has a low natural frequency of between 3 and2.5Hz.

When the air supply is switched off the block or plate will settle down and rest onpackers, this allows removal for maintenance (item 5 in Fig. 3.11). The air springsrequire a reliable, low flow, condensate free air supply that is not always easy toprovide to the lowest point in the cell system. Sufficient room must be left aroundthe bed plate to allow maintenance access to the air-spring units.

In cases where there is no advantage in having the bedplate face at ground level(as is required for pallet systems), the test stand bedplate can be mounted on the flatcell floor. It is possible that plant and engines mounted on rubber viscous mountsor air-spring systems could, unintentionally, become electrically isolated from theremainder to the facility by virtue of the rubber elements; it is vital that a commongrounding scheme is included in such facility designs (see Chapter 10, Electricaldesign considerations).

A special application concerns the use of seismic blocks for supporting enginesin anechoic cells. It is, in principle, good practice to mount engines undergoingnoise testing as rigidly as possible, since this reduces noise radiated as the result ofmovement of the whole engine on its mountings. Lowering of the centre of gravityis similarly helpful since the engines have to be mounted with a shaft centre heightof at least 1 metre to allow for microphone placement.

While the dynamometer is not a significant source of vibration, it is commonpractice to mount both engine and brake on a common block; if they are separatedthe relative movement between the two must be within the capacity of the connectingshaft and its guard.

Finally it should be remarked that there is available on the market a bewilderingarray of different designs of isolator or flexible mounting, based on steel springs,air springs, natural or synthetic rubber of widely differing properties used in com-pression or shear, and combinations of these materials. For the non-specialist themanufacturer’s advice should be sought and design sheets obtained (Fig. 3.12).

Summary of vibration section

This section should be read in conjunction with Chapter 9, which deals with theassociated problem of torsional vibrations of engine and dynamometer. The twoaspects – torsional vibration and other vibrations of the engine on its mountings –cannot be considered in complete isolation.

The exciting forces arising from (inevitable) unbalance in a reciprocating engineare considered and a description of the essential features of mounting design is given,together with a check list of points to be considered.


+180°

+90°

+180°

+90°

+270°

Displacement

Restoring force180° out of phase with displacement,urges towards centre: does not work

VelocityLeads displacement by 90°

AccelerationIn phase with restoring force

Driving forceIn phase with velocity: does work

Damping forceLags displacement by 90°

(or leads it by 270°): absorbs energy

+

Vibrations of a pendulum or spring and mass

–

Figure 3.12 Diagram of phase relationship of vibrations

36 Engine Testing

Noise: fundamentals

Sound intensity

The starting point in the definition of the various quantitative aspects of noisemeasurement is the concept of sound intensity, defined as:

I =p2

cW/m2

where p2 is the mean square value of the acoustic pressure, i.e. the pressure variationdue to the sound wave, the density of air and c the velocity of sound in air.

Intensity is measured in a scale of decibels (dB):

dB= 10 log10

(

I

I0

)

= 20 log10

(

p

p0

)

where I0 corresponds to the average lower threshold of audibility, taken by conventionas I0 = 10−12W/m2, an extremely low rate of energy propagation.

From these definitions it is easily shown that a doubling of the sound intensitycorresponds to an increase of about 3 dB (log10 2= 0.301). A tenfold increase givesan increase of 10 dB, while an increase of 30 dB corresponds to a factor of 1000in sound intensity. It will be apparent that intensity varies through an enormousrange.

The value on the decibel scale is often referred to as the sound pressure level

(SPL). In general, sound is propagated spherically from its source and the inversesquare law applies. Doubling the distance results in a reduction in SPL of about 6 dB(log10 4= 0.602).

The human ear is sensitive to frequencies in the range from roughly 16Hz to20 kHz, but the perceived level of a sound depends heavily on its frequency structure.The well-known Fletcher–Munson curves, Fig. 3.13, were obtained by averagingthe performance of a large number of subjects who were asked to decide when theapparent loudness of a pure tone was the same as that of a reference tone of frequency1 kHz.

Loudness is measured in a scale of phons, which is only identical with the decibelscale at the reference frequency. The decline in the sensitivity of the ear is greatestat low frequencies. Thus at 50Hz an SPL of nearly 60 dB is needed to create asensation of loudness of 30 phons.

Acoustic data are usually specified in frequency bands one octave wide. Thestandard mid-band frequencies are:

31.5 62.5 125 250 500 1000 2000 4000 8000 16 000Hz

e.g. the second octave spans 44–88Hz. The two outer octaves are rarely used innoise analysis.


Sound p

ressure

level (d

B)

(20

µP

α)

Equal loudness contours for pure tonesby octave (Fletcher–Munson curves)

110

100

90

80

70

60

50

40

30

20

10

0

31.5 63 125 250 500 1000 2000 4000 8000 16 000

Octave band centre frequency (Hz)

Thresholdof hearing

Figure 3.13 Fletcher–Munson curves of equal loudness

Noise measurements

Most instruments for measuring sound contain weighted networks which give aresponse to frequency which approximates to the Fletcher–Munson curves. In otherwords, their response to frequency is a reciprocal of the Fletcher–Munson relationship(Fig. 3.13). For most applications the A-weighting curve gives satisfactory resultsand the corresponding SPL readings are given in dBA. B- and C-weightings aresometimes used for high sound levels, while a special D-weighting is used primarilyfor aircraft noise measurements, Fig. 3.14.

The dBA value gives a general ‘feel’ for the intensity and discomfort level of anoise, but for analytical work the unweighted results should be used. The simplest typeof sound level meter for diagnostic work is the octave band analyser. This instrument

38 Engine Testing

C

B

A

A

B and C

5

0

–5

–10

–15

–20

–25

–30

–35

–40

–45

–50

Frequency (kHz)

Rela

tive r

esponse (

dB

)

0.02 0.05 0.1 0.2 0.5 1 2 5 10 20

Figure 3.14 Noise weighting curves

can provide flat or A-weighted indications of SPL for each octave in the standardrange. For more detailed study of noise emissions an instrument capable of analysisin one-third octave bands is more effective. With such an instrument it may, forexample, be possible to pinpoint a particular pair of gears as a noise source.

For serious development work on engines, transmissions or vehicle bodies muchmore detailed analysis of noise emissions is provided by the discrete Fourier transform(DFT) or fast Fourier transform (FFT) digital spectrum analyser. The mathematicson which the operation of these instruments is based is somewhat complex, but for-tunately they may be used effectively without a detailed understanding of the theoryinvolved. It is well known that any periodic function, such as the cyclic variationof torque in an internal combustion engine, may be resolved into a fundamentalfrequency and a series of harmonics (see Chapter 8, Dynamometers and the mea-surement of torque). General noise from an engine or transmission does not repeatin this way and it is accordingly necessary to record a sample of the noise over afinite interval of time and to process these data to give a spectrum of SPL againstfrequency. The Fourier transform algorithm allows this to be done.

Permitted levels of noise

Noise levels actually within an engine test cell nearly always exceed the levels per-mitted by statutes, while the control room noise level must be kept under observationand appropriate measures taken.

Member European states were required to bring into force the laws, regulationsand administrative provisions necessary to comply with the Physical Agents Directivebefore 15 February 2006. This directive lays down the minimum requirements forthe protection of workers from risks including to their health arising or likely to arisefrom exposure to noise and in a particular risk of permanent damage.


The Directive places the first action daily personal exposure level at 80 dBA(Noise at Work Act 85 dBA in 2005) and the second action level at 85 dBA (Noise ofWork Act 90 dBA in 2005). The employer’s obligations under prevailing regulationsinclude the undertaking of an assessment when any employee is likely to be exposedto the first action level. This means that hearing protection equipment must beavailable to all staff having to enter a running cell, where sound levels will oftenexceed 85 dBA, and that control room noise levels should not exceed 80 dBA.

Noise cancelling earphones, in place of the noise attenuating type, should beconsidered for use in high noise areas where vocal communication between staff isrequired.

Noise external to test facility and planning regulations

Almost all planning authorities will impose some form of restriction on the buildingof a new engine test laboratory that will seek to restrict the noise pollution caused.In the case of an existing industrial site, these restrictions will often take the formof banning any increase in sound levels at the nominated boundaries of the site. It isvital therefore that in the early planning and design stages a set of boundary noisereadings are taken by both the system integrator and the user. The survey should bebased on readings taken at important points, such as the boundary position nearestto residential buildings and located on an authorized map or with GPS readings.The readings should be taken in as normal working situation as possible withoutexceptional occurrences and at several times throughout the period the facility isto be used. Such a datum set of readings should be agreed as relevant with anyinterested party taken before work starts and will provide a key reference for anydispute post-commissioning.

Noise in the test cell environment

The measured value of SPL in an environment such as an engine test cell gives noinformation as to the power of the source: a noisy machine in a cell having goodsound-absorbent surfaces may generate the same SPL as a much quieter machinesurrounded by sound-reflective walls. The absorption coefficient is a measure ofthe sound power absorbed when a sound impinges once upon a surface. It is quitestrongly dependent on frequency and tends to fall as frequency falls below about500Hz.

Information on absorption coefficients for a wide range of structural materials andsound insulators is given in IHVE Guide B 12. A few approximate values are givenin Table 3.2 and indicate the highly reverberatory properties of untreated brick andconcrete.

40 Engine Testing

Table 3.2 Absorption coefficients at anoctave band centre frequency of 1 kHz

Concrete,brickwork

0�03

Glass 0�1Breeze blocks 0�5Acoustic tiles 0�9Open window 1�0

It should be remembered that the degree to which sound is absorbed by itssurroundings makes no difference to the intensity of the sound received directly fromthe engine.

‘Cross talk’ between test cell and control room and other adjacent rooms canoccur through any openings in the partition walls and also through air conditioningducts and other service pipes, when there is a common system.

Vehicle noise measurements

A number of regulations lay down permitted vehicle noise levels and the methods bywhich they are to be measured. In the EC it is specified that A-weighted SPL shouldbe measured during vehicle passage at a distance of 7.5m from the centre line of thevehicle path. The EEC rules on the permissible sound level and the exhaust systemof motor vehicles (70/157 adapted by 1999/101) covers all motor vehicles with amaximum speed of more than 25 km/h. The limits range from 74 dB(A) for motorcars to 80 dB(A) for high-powered goods vehicles.

Many motor sport venues worldwide have introduced their own noise limits forvehicles using their tracks, often in the range of 95–98 dBA measured at a fixedpercentage of engine speed and a set distance from the exhaust relying on the useof hand-held meters operated by scrutineers. Results of such tests can be highlymisleading if care is not taken to avoid sound reflection from surroundings, nor dothey presently measure noise levels coming from induction systems. Some tracks arenow fitted with complex ‘drive-by’ monitoring systems which more closely representthe ‘noise pollution’ created.

Anechoic test cells

The testing of engines and vehicles for noise, vibration and harshness (NVH) formsa substantial element of vehicle development programmes, and calls for special-ized (and very expensive) test facilities. Essentially, such a cell should provide anenvironment approximating as closely as possible to that of the vehicle on the road.


Such cells are of two kinds: semi-anechoic (in the USA, commonly and correctly‘hemi-anechoic’) in which walls and ceiling are lined with sound-absorbent materialswhile the floor is reflective, generally of concrete, and full anechoic cells, in whichall surfaces including the floor are sound absorbent.

A full anechoic cell is adapted to measuring the noise radiated in every directionfrom a source, while the semi-anechoic cell simulates the situation where the sourceis located in the open, but resting on a reflective horizontal plane; clearly, the latteris appropriate for land vehicle testing.

The design of anechoic cells is a matter for the specialist, but BS 4196, whilespecifically concerned with methods of determining sound power levels of noisesources, gives some useful guidance on certain aspects of anechoic cell design,including

• desirable shape and volume of the cell;• desirable absorption coefficient of surfaces;• specification of absorptive treatment;• guidance regarding avoidance of unwanted sound reflections.

A key part of the technical specification of an anechoic cell is known as the ‘cut-off’ frequency which is the frequency below which the rate of sound level decayin the chamber no longer replicates the outdoors or, more precisely, a ‘free-field’environment. The cut-off frequency normally required for automotive anechoic cells(both engine and vehicle) is around 120Hz.

The acceptance test of the anechoic characteristic of the cell is carried out usinga broadband generator placed at the geometric centre of the cell. The level ofsound decay is measured by a microphone physically moved along chords tensionedbetween the noise source and the high level corners of the cell. The decay measuredis compared with the 6 dB distance doubling characteristic of a free-field. The taskof certification of an anechoic cell and equipment required is a specialist field.

There are certain other points that should be borne in mind by the non-specialistif he is required to take responsibility for setting up a test facility of this kind:

• The structure of an anechoic cell should be isolated from any environment inwhich noise is generated, such as production plant. Common ventilation systemsshould be avoided as should any other paths to facility spaces that act for noiseor vibration transmission.

• The internal volume of the building shell will be roughly twice that of the usablespace because of the space required for the acoustic lining. This lining often takesthe form of foam shapes which may protrude some 700mm from the supportingstructure, which may itself be some 300mm from the cell wall.

• Access doors and windows affect the acoustic performance adversely and must bekept to a minimum commensurate with safety. The acoustic lining covering exitsshould be coloured to make their location obvious. Windows should preferablybe avoided since glass has poor sound-absorbent properties.

42 Engine Testing

• Acoustic research makes use of highly sensitive instruments and correspond-ingly sensitive measurement channels. The design of the signal cabling must becarefully considered at the design stage to avoid interference and a multitude oftrailing leads in the cell.

• The ventilation of acoustic cells presents particular problems. In most ordinarytest work, the noise contributed by the ventilation system may be ignored but thisis not the case with acoustic cells, where the best solution is usually to dispensewith a forced draught fan and to draw in the air by way of suitable filters andsilencers, using only an induced draught fan. However, full power running inanechoic cells is normally of short duration and the precise control of inlet airtemperature is not so critical as in cells concerned with power measurement.

• Acoustic cell linings usually absorb liquids and may be inflammable, increasingfire hazards. Particular attention should be paid to smoke detection and to thefire quenching system.

• The dynamometer must clearly be isolated acoustically from the cell proper. Forwork on engines and transmissions, the dynamometer is located outside the cell,calling for an overlong coupling shaft, the design of which will present problems.In many designs the shaft system is split into two sections with an intermediatebearing mounted on a support block just inside the anechoic chamber.

• For work on complete vehicles a rolling road dynamometer, Chapter 18, will bethe usual solution: this will transmit a certain amount of noise to the cell whichmay be measured by motoring the rolls in the absence of a vehicle.

• The engine in an anechoic cell must be raised above floor level, typically by1.0–1.2m to shaft centre line, to permit microphones to be located below it. Thiscalls for non-standard engine mounting arrangements.

Exhaust noise

The noise from test cell exhaust systems can travel considerable distances and bethe subject of complaints from neighbouring premises, particularly if running takesplace at night or during weekends. The design of test cell exhaust systems is largelydictated by the requirement that the performance of engines under test should not beadversely affected and by restraints on noise emission, both internally and externalto the building. Essentially, there are two types of device for reducing the noiselevel in ducts: resonators and absorption mufflers. A resonator, sometimes known asa reactive muffler, is shown in Fig. 3.15. It consists of a cylindrical vessel dividedby partitions into two or more compartments. The exhaust gas travels through theresonator by way of perforated pipes which themselves help to dissipate noise. Thedevice is designed to give a degree of attenuation, which may reach 50 dB, over arange of frequencies, see Fig. 3.16 as a typical example.

Absorption mufflers, which are the type most commonly used in test facilities,consist essentially of a chamber lined with sound-absorbent material through whichthe exhaust gases are passed in a perforated pipe. Absorption mufflers give broadband


Figure 3.15 Exhaust resonator or reactive muffler

damping but are less effective than resonators in the low-frequency range. However,they offer less resistance to flow.

Selection of the most suitable designs for a given situation is a matter for thespecialist. Both types of silencer are subject to corrosion if not run at a temperatureabove the dew point of the exhaust gas and condensation in an absorption muffler isparticularly to be avoided.

As with a number of engine attached devices, the exhaust silencer may be con-sidered as part of the engine rigging when it is engine model specific, or it may beconsidered as part of the cell, or both may be used. Modern practice tends to usethe vehicle exhaust system complete with silencers within the cell to extract the gasmixed with a proportion of cell air into a duct fitted, before exiting the building, viaabsorption attenuators. Fans used in the extraction are rated to work up to gas/airtemperatures of 200�C and the effect of the temperature on the density of the mixtureand therefore the fan power must be remembered in the design process.

Tail pipes

The noise from the final pipe section is directional and therefore often points sky-wards, although the ideal would be to terminate in a wide radius 90� bend away

44 Engine Testing

60

50

Attenuation (

dB

)

40

30

20

10

031.5 63 125 250 500

One third octave band centre frequency

Cut-off frequencyNoise at lower frequencies will be amplified

Performancepoor in pass bandregion

Pass band

1000 2000 4000 8000 16 000

Figure 3.16 Sample performance of a reactive muffler

from the prevailing wind. The pipe end can, if incorrectly positioned and with aplain 90� end, suffer from wind-induced pressure effects; this has led to difficultiesin getting correlation of results between cells. It has to be remembered that the con-densate of exhaust gases is very corrosive, also that rain, snow, etc., should not beallowed to run into undrainable catch points or the engine. Where there is a need tominimize any smoke plume, a tail pipe of the design shown in Fig. 6.7a may assistin mixing and dispersal.

Notation

Vibration

Mass of piston +1/3 connecting rod mp kgAngular velocity of crankshaft �cS

−1

Crank radius r mCrank angle from t.d.c.

Connecting rod length/r n


Unbalanced exciting force f NMass of engine m kgCombined stiffness of mountings k N/mAmplitude of vibration ×mAngular velocity of vibration � rad/sAngular velocity at resonance �0 rad/sNatural frequency n0 HzPhase angle A radDamping coefficient cN s/mDamping ratio C

Acceleration due to gravity g m/s2

Notation

r.m.s. value of acoustic pressure p N/m2

Density of air kg/m3

Velocity of sound in air c m/sSound intensity I W/m2

Threshold sound intensity Io W/m2

References

1. Ker-Wilson, W. (1959) Vibration Engineering, Griffin, London.2. Thompson, W.T. (1988) Theory of Vibration, 3rd edn, Prentice-Hall, London.3. Fader, B. (1981) Industrial Noise Control, Wiley, Chichester.4. Turner, J.D. and Pretlove, A.J. (1991) Acoustics for Engineers, Macmillan Education,

London.5. Warring, R.H. (1983) Handbook of Noise and Vibration Control, 5th edn, Trade and

Technical Press, Morden, UK.6. Maw, A.N. (1992) The design of resilient mounting systems to control machinery noise in

buildings. Plastics, Rubber and Composites Processing and Applications 18, 9–16.

Further reading

BS 3425 Method for the Measurement of Noise emitted by Motor Vehicles.

BS 3539 Specification for Sound Level Meters for the Measurement of Noise Emitted

by Motor Vehicles.

BS 4196 Parts 0 to 8 Sound Power Levels of Noise Sources.

BS 2475 Specification for Octave and One-third Octave Band-pass Filters.

BS 3045 Method of Expression of Physical and Subjective Magnitudes of Sound or

Noise in Air.

BS 4198 Method for Calculating Loudness.

46 Engine Testing

BS 4675 Parts 1 and 2 Mechanical Vibration in Rotating Machinery.

BS 5330 Method of Testing for Estimating the Risk of Hearing Handicap due to

Noise Exposure.

I.H.V.E. Guide B12: Sound Control, Chartered Institution of Building Services,London.

Noise at Work Regulations (1990) Health and Safety Executive, London.

4 Test cell and control room design:an overall view

Introduction

This chapter is intended for readers who may lack experience of engine testing ortest engineers having to consider the specification of test cells and deals in broadoutline with salient features of the main types of engine test cell and the associatedoperators control space, ranging from the simplest possible area for the user with anoccasional requirement to run a test, to complex multiple cell installations of vehiclemanufacturer and research organizations. Questions involved in the sizing of testcells, control rooms, the provision of services, engine handling and safety and firesuppression are discussed. Following a description of typical cell types, the subjectof optimum space required for the cell, control room and services is covered.

It is assumed that those involved in the design process have an operationalspecification (see Chapter 1) on which to base their discussions.

Overall size of individual test cells

One of the early considerations in planning a new test facility will be the spacerequired. The areas to be separately considered are

• the engine or powertrain test cell;• the control room;• the space required for services and support equipment;• the support workshop or engine rig and derig area;• the storage area required for engine rig items and consumables.

A cramped cell, in which there is not room to move around in comfort, is a permanentsource of danger and inconvenience. The smaller the volume of the cell the moredifficult it is to control the ventilation system under conditions of varying load (seeChapter 5). As a rule of thumb, there should be an unobstructed walkway 1 metrewide all round the test engine. Cell height is determined by a number of factorsincluding the provision or not of a crane beam in the structure. In practice, mostmodern automotive cells are between 4 and 4.5 metres internal height.

48 Engine Testing

Table 4.1 Some actual cell dimensions found in UK industry

6.5m long × 4m wide × 4m high QA test cell for small automotive dieselsfitted with eddy-current dynamometer

7.8m long × 6m wide × 4.5m high ECU development cell rated for 250 kWengines, containing work bench andsome emission equipment

6.7m long × 6.4m wide × 4.7mhigh

Gasoline engine development cell witha.c. dynamometer, special coolant andintercooling conditioning

9.0m wide × 6m × 4.2m high(to suspended ceiling)

Engine and gearbox development bedwith two dynamometers in ‘Tee’configuration. Control room runsalong 9.0 metre wall

It is often necessary, when testing vehicle engines, to accommodate the exhaustsystem as used on the vehicle, and this may call for space and extra length in thecell. It must be remembered that much of the plant in the cell requires calibrationfrom time to time and there must be adequate access for the calibration engineer, hisinstruments and in some cases ladder. The major layout problem may be caused bythe calibration of the dynamometer, involving accommodation of a torque arm anddead weights.

Control room space depends significantly upon the individual operations manage-ment; some are spaces where the engine operator stands and others are work spaceand conference area for a development team (Table 4.1).

Some typical test cell designs

The basic minimum

There are many situations in which there is a requirement to run engines under load,but so infrequently that there is no economic justification for building a permanenttest cell. Examples are organizations concerned with engine overhaul and rebuilding,and specialist engine tuners.

To meet needs of this sort all that is required is a suitable area provided with

• water supplies and drains;• portable fuel supply system;• adequate ventilation;• arrangements to take engine exhaust to exterior;• minimum necessary sound insulation;• adequate fire and safety precautions.

Test cell and control room design: an overall view 49

Figure 4.1 Portable engine test stand

Figure 4.1 shows a typical installation, consisting of the following elements:

• portable test stand for engine and dynamometer;• engine cooling system;• engine exhaust system;• control console.

A ‘bolt-on’ dynamometer (see Chapter 8) requires no independent foundation, andis particularly suitable for a simple mobile installation. This type of dynamometer isbolted to the engine bell housing, using an adaptor plate, with a splined shaft con-nection engaging the clutch plate. Occasionally, the dynamometer may be installedwithout removal of the engine from a truck chassis by dropping the propeller shaftand mounting the dynamometer on a hanger frame bolted to the chassis.

The dynamometer cooling water can be simply run from mains supply and run towaste or it can be circulated by a pump within a sump into which the water drains.The engine cooling system includes a cooling column that maintains the top-hosetemperature to a set figure and minimizes the quantity of cooling water run to wasteor back to the sump. In some cases the radiator associated with the engine under testand placed in front of a constant speed fan may be used.

The control console requires, as a minimum, indicators for dynamometer torqueand speed. The adjustment of the dynamometer flow control valve and engine throttlemay be by cable linkage or simple electrical actuators. The console should also house

50 Engine Testing

the engine stop control and an oil pressure gauge. A simple manually operated fuelconsumption gauge of the ‘burette’ type is adequate for this type of installation.

Typical applications are

• proving of truck or bus engines after overhaul or rebuild;• tuning of engines for racing in same workshops, etc.;• military vehicle overhaul in portable workshops;• checking, post-rebuild, emissions against legislative requirements.

General purpose automotive engine test cells, 50 to 450 kW

This category represents by far the largest number of test cells in service (see Figs 4.2and 4.3 for variants of this common layout).

Figure 4.2 General purpose test cell arranged against an outside wall withcontrol desk ‘side on’ to engine


EngineEngine

Double

Control

Room

Figure 4.3 Engine test cells with shared control rooms running down the lengthof the cell

Such cells are often built in multiples, side-by-side and in a line with a commoncontrol room that is shared by two cells on either side. Engines enter the cell byway of a large door in the rear wall while the operator may enter by way of adoor in the front wall to one side of the control desk. Typically there is a double-skinned toughened glass window in front of the control desk. Most wall-mountedinstrumentation, smoke meters, fuel consumption meters, etc., is carried on the sidewall remote from the cell access door.

In cells rated at above about 150 kW where engine changes and rigging werecarried out in the cell it was usual to provide a crane rail located above the testbed axis with a hoist of sufficient capacity to handle engines and dynamometer. Thepenalty of such a lifting beam is that the structure has to be built to take the full ratedcrane load plus its plant support load. In modern cells where the engine is riggedoutside the cell and trolley or pallet mounted, in-cell cranes are not usually includedas the cost benefit of a crane structure may be judged as marginal.

Figure 4.4 shows a dynamometer and engine support mounted on a common castiron tee-slotted bedplate; an assembly following this basic format forms the ‘island’at the centre of most engine test cells.

Signal conditioning units for the various engine transducers (pressure, temperature,etc.) are usually housed in a box carried by an adjustable boom cantilevered andhinged from the cell wall. The boom provides a route for segregated cable routesfrom engine instrumentation to the control room.

Boom boxes can be of considerable size, with contents that increase in the processof cell development. The ability of the wall to take the stresses imposed by the

52 Engine Testing

Figure 4.4 Dynamometer and engine mounting stand on a common tee-slotbedplate. This type of stand, requiring crane loading of the engine, is beingreplaced by pallet or trolley mounting systems

cantilevered load of the boom and its attachments should be carefully checked. Whereheadroom above the boom is restricted and it is not possible to fit a sufficientlyangled tie-wire then the boom may have to be attached directly to a steel pillar fixedat floor and ceiling.

In the cell illustrated (Fig. 4.2), exhaust pipes lead from the engine manifold to thein-cell exhaust and silencing system via a flexible connection section. Each exhaustoutlet can include an exhaust back pressure control valve.

Such a general purpose cell should be adaptable to take a wide range of engineswithin its thermal rating, but the number of changes is typically not more than oneor two per week. An adjustable engine mounting as shown in Fig. 4.4 is com-mon; such a universal stand can give maximum flexibility but set-up times may berather long.

The layout shown in Figs 4.2 and 4.3, with the test bed axis coinciding with thecontrol desk, gives good control window visibility, but is less easy to stack in amultiple cell layout than that of Fig. 4.5 which has the engine access and controlarea at the engine end of the cell. This layout works best when the control desk iswithin a wide corridor, but compromises the layout of the control space and makesconfidentiality of individual control areas difficult to arrange.


EngineEngine

Controldesk

Outside wall

Figure 4.5 Engine test cells having engine entry from the control corridor allow-ing side-by-side ‘stacking’ of cells


• development departments of engine and component manufacturers;• independent testing and development laboratories;• fuel and lubricant development and standard test procedures;• training and educational facilities;• military workshops;• specialized test cells for research and development.

Research and development engine and powertrain test cells

In this book, the term ‘powertrain’ testing is used to cover engine plus direct mountedtransmission, but does not cover vehicle transmissions complete with vehicle shafting;these are considered as ‘transmission test rigs’.

Powertrain rigs are often designed to be able to take up different configurations, ona large bedplate, as required by the unit under test (UUT). The typical configurationscatered for are

• transverse engine plus gearbox driving two dynamometers;• in-line engine and gearbox driving one dynamometer;• in-line engine plus gearbox driving two dynamometers (‘Tee’ set-up).

54 Engine Testing

To allow fast transition times, the various UUT have to be pallet mounted in a systemthat presents a common height and alignment to the cell interface points.

Automotive engine test and development facilities built since the mid-1990s willhave the following features, additional to those found in general purpose cells:

1. Exhaust gas analysis equipment (see Chapter 16)2. Dedicated combustion air treatment plant (see Chapter 6)3. Ability to run ‘in vehicle’ exhaust systems4. High dynamic four-quadrant dynamometers (see Chapter 8).

Such requirements increase both the volume of the cell and of the space requiredto house plant such as combustion air treatment equipment and the electrical drivecabinets associated with the a.c. dynamometers. The positioning and condition inwhich such ‘service’ plant operates is as important and as demanding as those of thecontents of the test cell.

While plant such as the a.c. drive cabinets are usually positioned outside thecell, for reasons of ambient temperature and noise, they need to be as close to thedynamometer as practical to avoid higher than necessary costs for the connectingpower cable.

Modern dynamometer drive cabinets tend to be large, heavy and difficult tomanoeuvre within restricted building spaces. Therefore, expert planning is required,as is anti-condensation heating if there is a long period between installation andcommissioning. Combustion air units also need to be close to the engine to preventheat loss or gain through the delivery trunking. They are often positioned in theservice room immediately above the engine, with the delivery duct running via a firedamper through the cell ceiling to a flexible duct attached to the engine as part ofthe rigging process. If the humidity of the combustion air is being controlled, theunit will require condensation and steam drains out of the building or into foul waterdrains.


• development departments of vehicle and engine manufacturers and major oilcompanies;

• motor-sport developers;• specialist consultancy companies;• government testing and monitoring laboratories.

Inclined, static or dynamic automotive engine test beds

With the increased use of ‘off-road’ vehicles there is likely to be an occasionalrequirement for test beds capable of handling engines running with the crankshaftcentre line inclined to the horizontal. This will present problems with hydraulicdynamometers with open water outlet connections. Closed circuit cooling systems, asare usual with eddy current machines, or electrical motor-based machines are easilyadapted to inclined running.


Some rigs have been built where the whole engine and dynamometer bedplateis mounted upon a system of hydraulic cylinders allowing for dynamic movementin three planes while the engine is running. The cost and operational complexity ofsuch installations will be obvious to most readers.

For the very occasional case of the engine with vertical crankshaft, e.g. outboardboat engines tested without dummy transmissions, the electrical dynamometer is theobvious choice and may generally be used without modification, although dry gapeddy current machines have also been used. Special arrangements need to be madefor torque calibration.

Automotive engine production test cells (hot test)

These cells are highly specialized installations forming part of an automation systemlying outside the scope of this book. The objective is to check, in the minimumpossible process time, that the engine is complete and runs. Typical ‘floor-to-floor’times for small automotive engines range between 5 and 8 minutes.

The whole procedure – engine handling, rigging, clamping, filling, starting, drain-ing and the actual test sequence – is highly automated, with interventions, if any,by the operator limited to dealing with fault identification. Leak detection may bedifficult in the confines of a hot test stand therefore it is often carried out at a special(black-light) station following test.

The test cell is designed to read from identity codes on the engine and recognizevariants to adjust the pass or fail criteria accordingly.

Typical measurements made during a production test include

• time taken for engine to start;• cranking torque;• time taken for oil pressure to reach normal level;• exhaust gas composition.

Most gasoline engines are no longer loaded by any form of dynamometer duringa hot test, but in the case with diesel engines load is applied with power outputmeasured and recorded.

Cold testing in production

In addition to rotational testing ‘in process’ of subassemblies, cold testing is some-times applied to (near) completely built engines. This is a highly automated process(see Chapter 19, section Data processing in cold testing).

Cold test rigs are invariably situated within the assembly process line fed bya ‘power and free’ or similar conveyor system. The mechanical layout, based ondocking on to a drive motor and transducer pick-up system, is physically simpler thana hot test cell since it does not require fuel supplies or hot gas and fluid evacuation.

56 Engine Testing

Cold test areas also have the cost advantage of being without any significant enclosureother than safety guarding.

Production hot test cell layout

In the design phase of a production hot test facility a number of fundamental decisionshave to be made, including

• layout, e.g. conveyor loop with workstations, carousel;• what remedial work, if any, to be carried out on test stand;• processing of engines requiring minor/major rectification;• engine handling system, e.g. bench height, conveyor and pallets, ‘J’ hook con-

veyor, automated guided vehicle;• engines rigged and derigged at test stand or remotely;• storage and recycling of rigging items;• maintenance facilities and system fault detection;• measurements to be made, handling and storage of data.

Production testing imposes heavy wear and tear, particularly on engine riggingcomponents which need constant monitoring and spares available.

Automatic shaft docking systems may represent a particularly difficult designproblem where multiple engine types are tested, and where faulty engines are crankedor run for periods giving rise to unusual torsional vibration and torque reversals.

Shaft docking splines need adequate lubrication and, like any automatic dockingitem, can become a maintenance liability if one damaged component is allowed totravel round the system, causing consequential damage to mating parts.

Modular construction of key subassemblies will allow repairs to be carried outquickly by replacement of complete units thus minimizing production downtime.

Large- and medium-speed diesel engine test areas

As the physical size of an engine increases the logistics of handling them becomesmore significant, therefore the test area is often located within the production plantclose to the final build area.

Above a certain size engines are tested within an open shop in which they havebeen finally assembled. The dynamometers designed to test engines in the rangesof about ≥20MW are small in comparison to the prime mover, therefore the testequipment is brought to the engine rather than the more usual arrangement wherethe engine is taken to a cell.

Cells for testing medium-speed diesels require access platforms along the sidesof the engine to enable rigging of the engine and inspection of the top mountedequipment, including turbochargers, during test. There is a design temptation toinstall services under these platforms but these spaces can be difficult and unpleasant


to access, therefore the maintenance items such as control valves should, wherepossible, be wall or boom mounted.

Rig items can be heavy and unwieldy. Rigging of engines of this size is adesign exercise in itself but the common technique is to prerig engines of differingconfigurations in such a way that they present a common interface when put in thecell. This allows the cell to be designed with permanently installed semi-automatedor power-assisted devices to connect exhausts, intercooler and engine coolant piping.The storage of rig adaptors will need careful layout in the rig/derig area.

Shaft connection is usually manual with some form of assisted shaft lift andlocation system.

Special consideration should be given in these types of test areas to the draining,retention and disposal of liquid spills or wash-down fluids.

Large engine testing often exceeds the normal working day, therefore running atnight or weekends is not uncommon and may lead to complaints of exhaust noiseor smoke from residential areas nearby. Each cell or test area will have an exhaustsystem dedicated to a single engine; traditionally and successfully the silencers havebeen of massive construction built from the ground at the rear of the cells. Modernversions may be fitted with smoke dilution cowls. These exhaust systems requirecondensate and rain drains to prevent accelerated corrosion.

Seeing and hearing the engine

Except in the case of production test beds, large diesel engines and portabledynamometer stands, it is almost universal practice to separate the control space fromthe cell proper. This space may be in a corridor joining several similar workstations.It can be a separate, cell specific, control room or shared and sandwiched betweenthe lengths of two cells.

Thanks to modern instrumentation and closed circuit television, a window betweenthe control room and test cell is rarely absolutely necessary, although many users willspecify one. The cell window causes a number of design problems; it compromisesthe fire rating and sound attenuation of the cell and uses valuable space in thecontrol room. The choice to have a window or not is often made on quite subjectivegrounds. Cells in the motor sport industry all tend to have large windows becausethey are visited by sponsors and press, whereas OEM multicell research facilitiesincreasingly rely on remote monitoring and closed-circuit television. With the adventof the thin, flat computer screen many cell windows have become compromised bybanked arrays of displays.

The importance or otherwise of engine visibility from a window is linked with afundamental question: which way round is the cell to be arranged?

It is often more convenient from the point of view of engine handling and instal-lation to have the engine at the rear, adjacent to a rear access door, but this gives theworst visibility if the design relies only on a window.

58 Engine Testing

The experienced operator will be concentrating attention on the indications ofinstruments and display screen, and will only catch changes happening in the cellout of the corner of an eye. It is important to avoid possible distractions, such asdangling labels or identity tags that can flutter in the ventilation wind.

Hearing has always been important to the experienced test engineer, who can oftendetect an incipient failure by ear well before it manifests itself in any other way.

Unfortunately, modern test cells, with their generally excellent sound insulation,cut off this source of information and consideration should be given to the provisionof in-cell microphones with external loudspeakers or earphones.

Flooring and subfloor construction

The floor, or seismic block when fitted (see Chapter 3, Vibration and noise), mustbe provided with arrangements for bolting down the engine and dynamometer. Agood solution is precisely level and cast in two or more cast-iron T-slotted rails.The machined surfaces of these rails form the datum for all subsequent alignmentsand they must be set and levelled with great care. Any twist could lead to seriousdistortion. The use of fabricated box beams is a false economy.

Sometimes complete cast-iron floor slabs with multiple T-slots are used as thecell floor rather than as a spring supported floor plate within a pit; these tend to trapliquids and are highly sound-reflective. They can also be very slippery.

Floors should have a surface finish that does not become unduly slippery whenfluids have been spilt; special marine deck paints are recommended over the morecosmetically pleasing smooth finishes; however, any surface finish used must be ableto resist the fluids, including fuels, used in the cell.

It is good practice to provide floor channels on each side of the bed, as they areparticularly useful for running ‘cold’ services and drains in an uncluttered manner;however, regulations in most countries call for spaces below floor level to be scav-enged by the ventilating system to avoid any possibility of the build-up of explosivevapours. Fuel services should preferably not be run in floor trenches.

Floor channels should be covered with well-fitting chequer plates, not weighingmore than about 20 kg each, and provided with lifting holes. The plates can be cutas necessary to accommodate service connections.

Doors

Doors that meet the requirements of noise attenuation and fire containment areinevitably heavy and require more than normal effort to move them; this is a safetyconsideration to be kept in mind when designing the cell. Forced or induced ven-tilation fans can give rise to pressure differences across doors, possibly making itdangerous or impossible to open a large door. The recommended cell depression forventilation control is 50 Pa.


All test cell doors must be either on slides or be outward opening. There aredesigns of both sliding and hinged doors that are suspended and drop to seal in theclosed position.

Sliding doors have the disadvantage of creating ‘dead’ wall space when open.Doors should be provided with small observation windows and may be subject toregulations regarding the provision of exit signs.

The cell operator should give consideration to interlocking the cell doors withthe control system to prevent human access during chosen operating conditions. Acommon strategy is to force the engine into a ‘no load/idle’ state as soon as a dooris opened.

Walls and roof

Test cell walls are required to meet certain special demands in addition to thosenormally associated with an industrial building. They, or the frame within which theyare built, must support the load imposed by any crane installed in the cell, plus theweight of any equipment mounted on or suspended below the roof. They must be ofsufficient strength and suitable construction to support wall-mounted instrumentationcabinets, fuel systems and any equipment carried on booms cantilevered out fromthe walls. They should provide the necessary degree of sound attenuation and mustcomply with requirements regarding fire retention (usually a minimum of one hourcontainment).

High density building blocks provide good sound insulation which may beenhanced by filling the voids with dried casting sand, but this may cause difficultiesin creating wall penetrations after the original construction. Walls of whatever con-struction usually require some form of internal acoustic treatment, such as 50-mmthick sound absorbent panels, to reduce the level of reverberation in the cell. Suchpanels can be effective on walls and ceilings, even if some areas are left uncoveredfor the mounting of equipment. The alternative, easier to clean, option of havingceramic tiled walls is not uncommon, where test operators are not allowed in thecell, to be exposed to high noise levels, during engine running.

The choice of cell wall material has been complicated in recent years by theavailability of special construction panels made of sound absorbent material sand-wiched between metal sheets, of which the inner (cell) side is usually perforated.These 100-mm thick panels may be used with standardized structural steel framesfor the construction of test cells and, while not offering the same level of soundattenuation as dense block walls, give a quick and clean method of construction witha pleasing finished appearance; this is particularly useful when cells are to be builtin an existing building. However, proper planning is necessary if heavy equipment isto be fixed as internal ‘hard-mounting’ points must be built into the steel structure.

The roof of a test cell often has to support the services housed above which mayinclude large and heavy electrical cabinets. Modern construction techniques such asthe use of ‘rib-decking’ (Fig. 4.6), which consists of a corrugated metal ceiling that

60 Engine Testing

Concrete floordecking

Reinforced concretefloor poured onto

metal decking

Metal rib-decking ceilingsur face

Figure 4.6 Section through metal and concrete ‘rib-decking’ construction of acell roof

225

600

Figure 4.7 Section through a concrete plank which gives a quick and highstrength cell roof construction, but can be difficult to modify post-installation

provides the base of a reinforced concrete roof pouring in situ, are commonly used.An alternative is hollow core concrete planking (Fig. 4.7), but the internal voids inthis material mean that a substantial topping of concrete screed is required to obtaingood sound insulation.

If penetrations have to be made in the structure after installation, these systemsare not particularly easy to modify, therefore good preplanning and well-integrateddesign of services is important.

False, fire retardant, ceilings suspended from hangers fixed into the roof can befitted if the ‘industrial’ look of concrete or corrugated metal is unacceptable, but thishas to be shaped around all of the roof penetrating service ducts and pipes and cannot always be financially justified.


Lighting

The typical test cell ceiling may be cluttered with fire sprinkler systems, exhaust out-lets, ventilation ducting and a lifting beam. The lights are often the last consideration,but are of vital importance. They must be securely mounted so as not to move in theventilation ‘wind’ and give a high and even level of lighting without causing glarefrom the control room window. Unless special and unusual conditions or regulationsapply, cell lighting does not need to be explosion proof; however, lights may beworking in an atmosphere of soot-laden fumes and so must be sufficiently robust, beeasily cleaned and operate at moderate exposed surface temperatures.

The detailed design of a lighting system is a matter for the specialist. The ‘lumen’method of lighting design gives the average level of illumination on a working planefor a particular number of ‘luminaries’ (light sources) of specified power arranged ina symmetrical pattern. Factors such as the proportions of the room and the reflectivityof walls and ceiling are taken into account.

The unit of illumination in the International System of Units is known as thelux, in turn defined as a radiant power of one lumen per square metre. The unit ofluminous intensity of a (point) source is the candela, defined as a source which emitsone lumen per unit solid angle or ‘steradian’. The efficiency of light sources in termsof candela per watt varies widely, depending on the type of source and the spectrumof light that it emits.

The IES Code lays down recommended levels of illumination in lux for differentvisual tasks. A level of 500 lux in a horizontal plane 500mm above the cell floorshould be satisfactory for most cell work, but areas of deep shadow must be avoided.In special cells where in situ inspections take place, the lighting levels may bevariable between 500 and 1000 lux.

Emergency lighting with a battery life of at least one hour and an illumination levelin the range of 30 to 80 lux should be provided in both test cell and control room.

It is sometimes very useful to be able to turn off the cell lights from the controldesk, whether to watch for sparks or red-hot surfaces or simply to hide the interior ofthe cell from unauthorized eyes. An outlet for a ‘wandering’ 110V inspection lampshould be included in the layout.

Engine handling systems

A considerable number of connections must be made to any engine before testingcan proceed. These include the coupling shaft, fuel, cooling water, exhaust and awide range of transducers and instrumentation. It is obviously cost-effective whendealing with high throughput of small automotive units to carry out such work in aproperly equipped workshop rather than use expensive test cell space. Once riggedthe engines have to be transported to the cell.

The degree of complexity of the system adopted for transporting, installing andremoving the engine within the cell naturally depends on the frequency with which

62 Engine Testing

the engine is changed. In some research or lubricant test cells the engine is more orless a permanent fixture, but at the other extreme the test duration for each engine ina production cell will be measured in minutes, and the time taken to change enginesmust be cut to an absolute minimum.

There is a corresponding variation in the handling systems:

• Simple arrangements when engine changes are comparatively infrequent. Theengine is either craned into the cell and rigged in situ or mounted on a suitablestand such as that shown in Fig. 4.4 which is then lifted, by crane or forklift, into thecell. All connections aremade subsequently by skilled staffwithworkshop back-up.

• The prerigged engine is mounted on a wheeled trolley or truck manoeuvred palletcarrying various transducers and service connections which are coupled to theengine before it is moved into the cell. An engine rigging workshop needs tobe fitted with a dummy test cell station that presents datum connection pointsidentical in position and detail to those in the cells, particularly the dynamometershaft in order for the engine to be pre-aligned. Clearly to gain maximum benefit,each cell in the facility needs to be built with critical fixed interface items inidentical positions from a common datum. These positions should be repeated bydummy interface points in any prerigging stand, in the workshop.

• For production test beds it is usual to make all engine to pallet connections priorto the combined assembly entering the cell. An automatic docking system permitsall connections (including in some cases the driving shaft) to be made in secondsand the engine to be filled with liquids.

Workshop support in the provision of suitable fittings and adaptors should alwaysbe made available. All pipe connections should be flexible and exhaust connectionscan be particularly troublesome and short-lived at high temperatures; they should betreated as consumable items.

Rigging of exhaust systems often requires a welding bay; this must obey regu-lations concerning shielding. Electrical arc welding must not be allowed within thetest cell because of the danger of damage to instrumentation by stray currents.

The workshop area used for derigging should be designed to deal with theinevitable spilled fluids and engine wash activities. Floor drains should run into anoil intercept unit.

Cell support services space

The design criteria of the individual services are covered in other chapters but at theplanning stage suitable spaces have to be reserved for the following systems listedin order of space required:

• ventilation plant including fans, inlet and outlet louvers and ducting includingsound attenuation section (see Chapter 5);

• engine exhaust system (see Chapter 6);


• electrical power distribution cabinets, including large drive cabinets in the caseof a.c. dynamometers (see Chapter 10);

• fluid services including cooling water, chilled water, fuel and compressed air (seeChapters 6 and 7).

In addition to these standard services, space in some facilities needs to include

• combustion air treatment unit (see Chapter 5);• exhaust gas emissions equipment (see Chapter 16).

It is very common to mount these services above the cell on the roof slab and it isoften desirable, although rarely completely possible, for the services of individualcells to be contained within the footprint of the cell and control room below.

Control room design

As the operator will spend a great deal of time in the control room its layout shouldbe as convenient and comfortable as possible. There should be a working surfaceaway from the control station, where discussions may be held. A keyboard will beneeded, but not all the time and in production cells it should be kept in a drawerwhen not in use.

Multiple computers can lead to several identical cordless keyboards and ‘mice’mixed together on a dangerously confusing desk space.

Some equipment often installed in control rooms, such as emissions instrumen-tation, can have a quite large power consumption. This must be taken into accountwhen specifying the ventilation system, which should conform to office standards.

The control console itself calls for careful design on sound ergonomic principles,and should, initially, include space for later additions. Figure 4.2 shows a typicalgeneral purpose control desk consisting of a work surface and an angled 19-in rackinstrumentation tower. By definition, the primary purpose of the control desk is eitherto permit the operator to control the engine directly, or to allow him to manage andmonitor a control system that may be automated to a greater or less degree. Prideof place will thus be given to the controls that govern engine torque and speed, theprime variables, and to the corresponding indicators of these variables, which maytake various forms.

Of almost equal importance are secondary indications and warning signals cover-ing such features as coolant temperature, lubricating oil pressure, exhaust tempera-ture, etc.

Many modern systems are now integrated in a pedestal desk layout with thedesk space supporting multiple screen displays that display all the test and facilitycondition data required.

The size and position of the observation window greatly affects the control roomlayout. In special climatic or anechoic cells windows are not fitted because theycompromise the integrity of the structure and function.

64 Engine Testing

If a window is fitted, it will normally be supplied by a specialist contractor andconsist of two or three sheets of toughened glass, typically about 10 cm apart, in aconfiguration designed to reduce internal reflection.

Displays and controls for secondary equipment such as cell services can berelegated to lower levels in the cabinets, while instruments used only intermittently,such as fuel or smoke meters, should be installed towards the top and operated ifnecessary from a standing position. The positioning of graphic displays or computermonitors can be a problem, as they occupy considerable space and the solution maybe to use articulated support booms.

The choice of indicator depends on the type of control and the kind of work to bedone. If, as is sometimes the case, the operator needs to manage the engine manuallyby a T-handle or rotary dial type throttle control he should have an analogue speedindicator prominently in view.

In general the analogue indicator, in which a pointer moves over a graduateddial, is the appropriate choice when the quantity indicated is under manual control.If the indicator is required to show the stage reached in a test, current levels ofsuch quantities as power and speed and any departures from desired values, a digitalreadout, perhaps accompanied by warning signals, is appropriate.

Presentation of readings in the form of a row of coloured bars on a screen canbe useful if the balance of a number of inter-related quantities is to be kept underreview, as is sometimes the case in gas turbine testing.

The operator’s desk should be designed so that liquids spilled on the workingsurface cannot run down into electronic equipment mounted beneath. Control roomsshould have two emergency escape routes. The floors should have non-slip andanti-static surfaces.

Instruments and controls: planning the layout

Even if the aboveprinciples are kept firmly inmind, it requires clear designmanagementto prevent the test cell and control room from degenerating over time into a jumble ofdifferent devices positioned in available spaces. This is not only inefficient but unsafe.

Computerization has greatly complicated the problem. Before total computerdependency, it was unusual to see in test cells a direct display of more than per-haps 10 or 12 quantities with possibly 10 more on a multi-pen chart recorder. Incomputer-controlled cells, particularly those associated with engine mapping, it isnot unusual to have six display screens mounted on a frame, many of them givingrapidly changing indications.

This can present the control system designer with a major problem in seeing ‘thewood for the trees’ and organizing a coherent display. In most cases the suppliershave passed the problem to the user by allowing them freedom to produce their owncustomized display screens. The problem, while transferred, has not been solved;some form of optimization and discipline of data display are required within the userorganization if serious misunderstandings or operational errors are to be avoided.


It may be found helpful, in planning the layout of instruments and controls, toclassify them under the following headings:

• operational and safety instrumentation;• primary instrumentation;• secondary instrumentation.

If the status or output of most or all of the instrumentation is shown on one or moredisplay screens, these displays must be arranged so that those of high priority arevisually prominent.

Primary instrumentation

This covers the essential data relating to the experimental purpose of the cell. Thedisplay of this information will be arranged in a prominent position where it iscomfortably seen from the operator’s desk.

The definition of primary as opposed to secondary instrumentation dependsentirely on the purpose of the cell. In a cell intended for noise measurement studies,for example, sound level indicators and recorders have pride of place while indica-tions of engine torque, exhaust temperature or fuel consumption are of secondaryimportance.

A test cell devoted to diesel engine exhaust emissions has as primary instrumen-tation the required gas analysis system, with all the associated controls of the gasdilution and sampling system. If, as is often the case, the engine is being put througha computer-controlled transient drive cycle, information regarding engine state is ofsecondary importance and displayed accordingly.

Secondary instrumentation

Information not of direct relevance to the test should be displayed away from theoperator’s normal field of attention, or it may be recorded on selectable pages of themain VDU display. Typical secondary information concerns slowly changing quan-tities such as barometric pressure, relative humidity and cooling water temperature.

In-cell controls

There are a number of tasks that may require an operator to be in the cell while theengine is running. Examples are the tuning of racing car engines, the setting of idlingadjustments and checking for leaks. It is very desirable, for safety reasons, that theengine is brought via door interlock to idle when the cell door is opened and thatthere should always be a second operator at the control desk while a colleague iswithin the running cell.

A simple control station within the cell, enabling the operator there to controlthe engine and observe torque and speed, is often useful but must be interlocked to

66 Engine Testing

prevent control actions being taken outside the cell at the same time. Duplication ofall control desk displays within the cell is seldom justified.

Emergency stop and engine shutdown

The engine test cell control system must include a robust method of shutting downthe engine and dynamometer (see Chapter 11, under sections Safety systems andEmergency stop function). All test facilities must be provided with a number ofemergency stop buttons, of large size and conspicuously marked, both at the controldesk and at strategic positions in the cell. An emergency stop button is intended foruse in the case of a local emergency; it operates only in the cell affected.

Since spark ignition engines may be stopped by interrupting the supply to theignition circuit this is usually linked to the emergency stop system. A diesel enginecalls for a more robust shutdown system: while many automotive diesels are fittedwith a shutdown solenoid, which should be wired into the emergency stop system,it should be backed up by some means of independently moving the fuel pump rackto the stop position. Some test plant suppliers produce standard devices or modifiedthrottle controllers which perform this function, using stored pneumatic power as aprecaution against electrical failure.

There is always the possibility of a diesel engine running away, usually as a resultof being fuelled by lubricating oil: this puts it out of the control of the fuel pump.Such an occurrence is rare but not unknown, even with modern engines. Horizontalbus engines, if inadvertently overfilled with lubricating oil, are prone to this trouble.If the emergency stop system is activated when an engine is running away all loadis removed from electrical or eddy current dynamometers and the situation is madeworse.

The two ways of stopping the engine in this situation are either by closing offthe combustion air supply or applying sufficient load to stall the engine. Unless theengine has been specially rigged with a closure valve or a means of injecting CO2,the former method is often not practicable. Stalling the engine in such an emergencycalls for a cool head and is best arranged by providing a special ‘fast stop’ buttonwhich ramps the dynamometer to full torque and de-energizes the engine systems.Such rapid stalling of the engine can be damaging: a fast stop button should only belocated at the control desk under a protective flap and operators should be carefullybriefed on its use.

Fire control

This subject has two separate system design subjects:

• the detection of flammable, explosive or dangerous gases in the cell and theassociated actions and alarms;

• the suppression of fire and the associated actions and alarms.


The gas detection system, usually supplied by a specialist subcontractor, will belinked to the cell control system and building management system as required by theoperator and in conformity to local and national regulation.

The gas hazard alarm and fire extinguishing systems are always entirely indepen-dent of and hierarchical to the emergency stop arrangements. They are linked to theventilation control system since in the event of a fire, the cell needs to be isolatedfrom sources of air and the fire needs to be contained.

The operation of both systems should be specified within the control and safetyinterlock matrix mentioned in Chapter 10, Electrical design considerations.

There is much legislation relevant to industrial fire precautions and also a numberof British Standards. The Health and Safety Executive, acting through the FactoryInspectorate, is responsible for regulating such matters as fuel storage arrangements,and should be consulted, as should the local fire authority.

Safety regulations and European ATEX codes applied to engine test cells

Atmospheric Explosion (ATEX) regulations were introduced as part of harmoniza-tion of European regulations for such industries as mines and paper mills whereexplosive atmospheres occur; the engine and vehicle test industry was not explicitlyidentified within the wording, therefore the European industry has had to negotiatethe conditions under which test cells may be excluded. As with all matters relatingto regulation, it is important for the local and industrial sector specific interpretationto be checked.

The classification of zones is as shown in Table 4.2.As has been determined over many years, secondary explosion protection measures

such as using EX-rated equipment (even if it were available) in the engine test cellmakes little sense since the ignition source is invariably the engine itself. Thereforeit is necessary to use primary explosion prevention methods that prevent the spacefrom ever containing an explosive atmosphere covered by ATEX.

Table 4.2 ATEX and US zone classification for occurrence of explosiveatmospheres

USA divisions ATEX zone

designation

Explosive gas

atmosphere exists

Remark

Zone 0 Continuously, orfor long periods

>1000 h/year

Division 1= zone0 and 1

Zone 1 Occasionally 10–1000 h/year

Division 2= zone2

Zone 2 For a short periodonly

<10 h/year

68 Engine Testing

These primary precautions, which cover both gasoline- and diesel-fuelled bedswithout distinction, are certified in Europe by the relevant body TUF and are asfollows:

• The cell space must be sufficiently ventilated both by strategy and volume flowto avoid an explosive atmosphere.

• There has to be continuous monitoring and alarming of hydrocarbon concentration(normally ‘warning’ at 20 per cent of lower explosive mixture and ‘shutdown’ at40 per cent).

• Leak-proof fuel piping using fittings approved for use with the liquids contained.• The maximum volume of fuel ‘available’ in the cell in the case of an emergency

or alarm condition is 10 litres.

With these conditions fulfilled, the only EX-rated electrical devices that needto be included in the cell design are the gas detection devices and any purgeextraction fan.

The requirement for the purge fan, which may have a dual purge/ventilationextract role can be achieved by use of a copper shroud ring for the impeller whichprevents sparks in the case of a rotor/stator contact.

Fire extinguishing systems

In the period between the second and third edition of this book, some gas-based firesuppressant systems have been banned for use in new building systems having fallenfoul of environmental protection legislation.

Table 4.3 lists the common fire suppression technologies and summarizestheir characteristics which are then covered in more detail in the followingparagraphs.

Microfog water systems

Microfog, or high pressure mist systems, unlike other water-based fire extinguishingsystems, have the great advantage that they remove heat from the fire source and itssurroundings and thus reduce the risk of reignition when it is switched off.

Microfog systems use very small quantities of water and discharge it as a veryfine spray. They are particularly efficient in large cells, such as vehicle anechoicchambers, where the fire source is likely to be of small dimensions relative to thesize of the cell.

Other advantages of these high pressure mist systems are that they tend to entrainthe black smoke particles that are a feature of engine cell fires and prevent orconsiderably reduce the major clean-up of ceiling and walls. Such systems are usedin facilities containing computers and high-powered electrical drive cabinets withminimum damage after activation and fast restart.


Table 4.3 Characteristics of major fire suppression systems

Water

sprinkler

Inert gas

(CO2)

Chemical

gases

High

pressure

water mist

Cooling effecton firesource

Some None None Considerable

Effect onpersonnel incell

Wetting Hazardous/fatal Minor healthhazards

None

Effect onenvironment

Highvolumeofpollutedwater

Greenhouseand ozonelayerdepleting

Greenhouse andozone layerdepleting

None

Damage byextin-guishingagent

Waterdamage

None Possible corro-sive/hazardousbyproducts

Negligible

Warningalarm timebeforeactivation

Nonerequired

Essential Essential Nonerequired

Effect onelectricalequipment

Extensive Small Possiblecorrosive byproducts

Small

Oxygen dis-placement

None In entire cellspace

In entire cellspace

At fire source

Carbon dioxide

Carbon dioxide (CO2) was used extensively in the industry until the environmentalimpact was widely understood. While they can be used against flammable liquid fires,they are hazardous to life, therefore a warning alarm period must be given beforeactivation to ensure staff have evacuated the cell. Breathing difficulties becomeapparent above a concentration of 4 per cent and a concentration of 10 per cent canlead to unconsciousness or, after prolonged exposure, to death.

CO2 is about 1.5 times heavier than air and it will tend to settle at ground level inenclosed spaces. The discharge of a CO2 flood system is likely to be very noisy andproduce a sudden drop of cell temperature causing dense misting of the atmosphereto take place that obscures any remaining vision possible through smoke.

70 Engine Testing

Dry powder

Powders are designed for high-speed extinguishment of highly flammable liquidssuch as petrol, oils, paints and alcohol; they can also be used on electrical or enginefires. It must be remembered that dry chemical powder does not cool nor does ithave a lasting smothering effect and therefore care must be taken against reignition.

Halons

Following the Montreal Protocol, some halons (halogenated hydrocarbons) are beingphased out. These contain chlorine or bromine, thought to be damaging to theozone layer, and their production has been banned. Unfortunately, this ban embracesHalon 1211 and Halon 1301, both hitherto used for total flooding and in portableextinguishers for dealing with flammable liquid fires. It is still feasible to use existingstocks, but clearly there are strong arguments for seeking replacements. Some areavailable, mostly fluorinated hydrocarbons, but care must be exercised as they mayhave toxic effects at their design concentrations.

Inergen

Inergen is the trade name for the extinguishing gas mixture of composition 52 percent nitrogen, 40 per cent argon, 8 per cent carbon dioxide. There are other alternativefire suppressants, including pure argon, many of which may be used in automaticmode even when the compartment is occupied, provided the oxygen concentrationdoes not fall below 10 per cent and the space can be quickly evacuated.

With all gaseous systems, precautions should be taken to ensure that accidentalor malicious activation is not possible. In carbon dioxide systems, automatic modeshould only be used when the space is unoccupied.

Total flood systems, of whatever kind, are usable only after the area has beenevacuated and sealed. They must be interlocked with the doors and special warningsigns must be provided.

Foam

Foam extinguishers could be used on engine fires but they are more suited toflammable liquid spill fires or fires in containers of flammable liquid. If foam isapplied to the surface or subsurface of a flammable liquid it will form a protectivelayer. Some powders can also be used to provide rapid knock-down of the flame.Care must be taken, however, since some powders and foams are incompatible.

Warning: Wall penetrations between control space and cell should be sealed bythe use of Hawke Gland™ boxes or similar devices. Poor maintenance or incompleteclosure of these fire-breaks, commonly caused by frequent installation of temporarylooms, can allow fire or extinguishant such as CO2 to escape explosively.


Vapour detection

Explosion proof devices are available for detection of preset levels of oil mist,hydrocarbon vapour and the gases used in engine test cells. It is common to set thealarm system handling these devices with a two-stage setting since their use presentsa practical problem in setting the sensitivity to a level that warns of real dangerwithout continuously tripping at a level which experience shows to be safe. Onceagain, the strategy controlling the alarms and interlocks should form part of the alarmand interlock matrix within the functional specification (see Chapter 10, Electricaldesign considerations).

Summary

The potentially hazardous nature of the test cell environment is emphasized andemergency stop and fire precautions are described. There are a number of differentbasic types of test cell, depending on the use for which they are intended. Over- andunderprovision of facilities should both be avoided.

The special features of test cell construction, which differs from that of any otherkind of industrial building, are described and references given to chapters of thisbook covering the design of specific services.

References

1. http://www.planningportal.gov.uk/england/professionals/

5 Ventilation and air conditioning

Introduction

In this chapter, the strategies of test cell ventilation are reviewed and the conceptof the test cell as an open system is applied to the analysis of thermal loadings andventilation requirements.1�2 The design process for ventilation systems is describedwith a worked example.

It must be remembered that an internal combustion (IC) engine is essentially anair engine and that the air used by the engine may come from the cell ventilation airor from a treatment unit outside the cell. Whatever its source, the performance andpower output of the engine is affected by the condition, temperature, pressure andhumidity of the ingested air.

The purpose of air conditioning and ventilation is the maintenance of an acceptableenvironment in an enclosed space. This is a comparatively simple matter whereonly human activity is taking place, but becomes progressively more difficult asthe energy flows into and out of the space increase. An engine test cell representsperhaps the most demanding environment encountered in industry. Large amounts ofpower will be generated in a comparatively small space, surfaces at high temperatureare unavoidable and large flows of cooling water, air and electrical power have tobe accommodated, together with rapid variations in thermal load.

Test cell ventilation strategies

While much of this chapter covers the most commonly used method of removingengine-generated heat from the test cell, by forced ventilation using ambient (outside)air, it is possible to recirculate cell air through a conditioning system.

The final choice will be influenced by the range of ambient conditions, buildingspace restraints and the type of testing being carried out.

Some authorities still use ‘room volume changes per hour’ as a measure ofventilation requirements; in this book the measure is only used to deal with legislativeor safety guidelines such as in the following table.

Ventilation systems for test cells not only remove heat but also prevent the build-up of dangerous levels of gases and vapours. Such requirements are dealt with byspecifically designed vapour purge systems (see below) and by ensuring sufficientair flow through the cell. The flow required to remove the heat generated by an

Ventilation and air conditioning 73

Table 5.1 Suggested air changes per hour, ‘rule of thumb’

Banks 2–4 Laboratories 5–10Cafés 10–12 Offices 5–7Shower rooms 15–20 Toilets 6–10Garages 6–8 Factories 8–10

engine or vehicle usually provides sufficient air flow to clear gasoline vapour andexhaust gases under most normal circumstances and the correct control strategy willensure that the cell space runs slightly below ambient pressure thus preventing suchgases being pushed into other work areas.

Purge fans: safety requirements to reduce explosion risk

In cells using volatile fuels the ventilation system will have to incorporate a purgefan, the purpose of which is to remove heavier than air and potentially explosivefumes from the lowest points of the cell complex. The purge fan plays an essentialrole in reducing the fire and explosion risk in a test cell (for relevance to ATEXregulations, see Chapter 4). The purge system should be integrated into the test cellcontrol system in such a way as to ensure, on cell start-up, that it has run for aminimum of 10min, with no hydrocarbon sensors showing an alarm state, beforeengine ignition and therefore fuel inlet valves are allowed to be energized.

The rating of the purge fan may be covered by local legislation, but as a minimumshould provide for 30 air changes of the enclosed cell space per hour.

The purge suction duct should extend to the lowest point of the cell within anyservices trench system; in some designs the purge fan is also used for providingexhaust dilution. Purge fan flow is additive to the outgoing flow of the main extractfan which has to be balanced by the replenishment volume in balanced systems. Thesource of the replenishment may be the combined flow of the combustion air systemand the main inlet air fan or the main inlet fan alone. Whatever strategy is used, thecell should run, when the doors are closed, at a negative pressure of around 30–50Pa compared with ambient. Any greater than this will make it difficult to initiatedoor opening.

If doors of the cell open for more than a few seconds, then the cell can beconsidered as operating in ‘workshop mode’ with fuel system turned off, then it ispossible to balance the treated combustion air flow with that of the purge fan so asto provide a space heating system. In cells with no combustion air system a separate‘comfort’ mode air heating and cooling system may have to be provided for the cell.

Air handling units

The use of an air handling unit (AHU) capable of recirculation of the majority, orvariable proportion, of test cell air through a temperature conditioning system can

74 Engine Testing

have operational advantages over a forced ventilation system. They may occupyless building space and may greatly reduce the problem of engine noise break-out.Designs are usually based on packaged units, supplied with building services such aschilled and medium pressure hot water (MPHW) supply. They are normally mountedabove the cell in the services room and aligned on the long axis of the cell tominimize ducting. A more rare alternative arrangement, in high roofed cells, usesceiling-mounted units that are based on a direct evaporative cooling and electricalheating units.

Air handling units, particularly the direct evaporative type, can be energy efficientbut a detailed and site-specific calculation of initial and running costs is needed tomake a valid comparison. Since closed systems will recirculate entrained pollutants,a separate combustion air system or alternative make-up source will always berequired to prevent build-up and to balance the loss through the low level purge.

The heat capacity of cooling air

By definition, the test cell environment is mainly controlled by regulating the quantity,temperature and in some cases the humidity of the air passing through it.

Air is not the ideal heat transfer medium: it has low density and specific heat andis transparent to radiant heat, while its ability to cool hot surfaces is much inferiorto that of liquids.

The main properties of air of significance in air conditioning may be summarizedas follows:

the gas equation:

pa ×105 = �R�ta +273� (1)

where:

pa = atmospheric pressure, bar�= air density, kg/m3

R= gas constant for air = 287 J/kgKta = air temperature, �C

Under conditions typical of test cell operation, with ta = 25�C (77�F), and standardatmospheric pressure (see Units and conversion factors), the density of air, fromeq. (1), is:

�=1�01325×105

287×298= 1�185kg/m3

or about 1/850 that of water.


The specific heat at constant pressure of air at normal atmospheric conditions isapproximately:

Cp = 1�01kJ/kgK

or less than one quarter that of water.The air flow necessary to carry away 1 kW of power with a temperature rise of

10�C is:

m=1

1�01×10= 0�099kg/s= 0�084m3/s= �2�9 ft3/s�

This is a better basis for design than any rule of thumb regarding number of cell airchanges per hour.

Heat transfer from the engine

It is useful to gain a feel for the relative significance of the elements that make upthe total of heat transferred from a running engine to its surroundings by consideringrates of heat transfer from bodies of simplified form under test cell conditions.3

Consider a body of the shape sketched in Fig. 5.1. This might be regarded asroughly equivalent, in terms of projected surface areas in horizontal and verticaldirections, to a gasoline engine of perhaps 100 kW maximum power output, althoughthe total surface area of the engine could be much greater. Let us assume the surfacetemperature of the body to be 80�C and the temperature of the cell air and cellwalls 30�C.

Heat loss occurs as a result of two mechanisms: natural convection and radiation.The rate of heat loss by natural convection from a vertical surface in still air is givenapproximately by:

Qv = 1�9 �ts− ta�1�25W/m2 (2)

80°C

0.6

m

0.9 m0.7 m

Figure 5.1 Simplified model of exampled 100 kW engine, for analysis of heattransfer to surroundings

76 Engine Testing

The total area of the vertical surfaces in Fig. 5.1= 1.9m2. The corresponding con-vective loss is therefore:

1�9×1�9× �80−30�1�25 = 480W

The rate of heat loss from an upward facing horizontal surface is approximately:

Qh = 2�5 �ts− ta�1�25W/m2 (3)

giving in the present case a convective loss of:

0�63×2�5 �80−30�1�25 = 210W

The heat loss from a downward facing surface is about half that for the upwardfacing case, giving a loss = 110W.

We thus arrive at a rough estimate of convective loss of 800W. For a surfacetemperature of 100�C, this would increase to about 1200W.

However, this is the heat loss in still air: the air in an engine test cell is anythingbut still and very much greater rates of heat loss can and probably will occur. Thiseffect must never be forgotten in considering cooling problems in a test cell: anincrease in air velocity greatly increases the rate of heat transfer to the air and maythus aggravate the problem.

As a rough guide, doubling the velocity of air flow past a hot surface increasesthe heat loss by about 50 per cent. The air velocity due to natural convection in ourexample is about 0.3m/s.

An air velocity of 3m/s would be moderate for a test cell with ventilating fansproducing a vigorous circulation, and such a velocity past the body of Fig. 5.1 wouldincrease convective heat loss four-fold, to about 3.2 kW at 80�C and 4.8 kW at 100�C.

The rate of heat loss by radiation from a surface depends on the emissivity ofthe surface (the ratio of the energy emitted to that emitted by a so-called blackbody of the same dimensions and temperature) and on the temperature differencebetween the body and its surroundings. Air is essentially transparent to radiation,which thus serves mainly to heat up the surfaces of the surrounding cell; this heatmust subsequently be transferred to the cooling air by convection, or conducted tothe surroundings of the cell.

Heat transfer by radiation is described by the Stefan–Boltzmann equation, a formof which is:

Qr = 5�77�

[

(

ts+273

100

)4

−

(

tw+273

100

)4]

(4)

A typical value of emissivity (�) for machinery surfaces would be= 0.9,ts = temperature of hot body (�C) and tw = temperature of enclosing surface (�C).


In the present case:

Qr = 5�77×0�9

[

(

353

100

)4

−(

303

100

)4]

= 370W/m2

In our present example, total surface area = 3.16m2, giving a radiation heat loss of1170W.

For a surface temperature of 100�C, this would increase to 1800W.

Heat transfer from the exhaust system

The other main source of heat loss associated with the engine is the exhaust system.In the case of turbocharged engines this can be particularly significant. Assume inthe present example that exhaust manifold and exposed exhaust pipe are equivalentto a cylinder 80mm diameter × 1.2m long at a temperature of 600�C, surface area0.3m2. Heat loss at this high temperature will be predominantly by radiation andequal to:

0�3×5�77×0�9

[

(

873

100

)4

−

(

303

100

)4]

= 8900W

from eq. (4).Convective loss is:

0�3×1�9× �600−30�1�25 = 1600W

from eq. (2).It is clear that this can heavily outweigh the losses from the engine and points to

the importance of reducing the run of unlagged exhaust pipe as much as possible.Heat losses from the bodies sketched in Fig. 5.1 and described above representing

an engine and exhaust system, surroundings at 30�C, may be summarized as follows:

Convection Radiation

engine, jacket and crankcase at 80�C 3.2 kW 1.2 kWengine, jacket and crankcase at 100�C 4.8 kW 1.8 kWexhaust manifold and tailpipe at 600�C 1.6 kW 8.9 kW

Heat transfer from walls

Most of the heat radiated from engine and exhaust system will be absorbed by the cellwalls and ceiling, also by instrument cabinets and control boxes, and subsequentlytransferred to the ventilation air by convection.

78 Engine Testing

4 m

6 m5 m

Figure 5.2 Simplified test cell for heat transfer calculation

Imagine the ‘engine’ and ‘manifold’ considered above to be installed in a test cellof the dimensions shown in Fig. 5.2. The total wall area is 88m2. Assuming a walltemperature 10�C higher than the mean air temperature in the cell and an air velocityof 3m/s the rate of heat transfer from wall to air is in the region of 100W/m2, or8.8 kW for the whole wall surface, roughly 90 per cent of the heat radiated fromengine and exhaust system; the equilibrium wall temperature is perhaps 15�C higherthan that of the air.

While an attempt to make a detailed analysis on these lines, using exact values ofsurface areas and temperatures, would not be worthwhile, this simplified treatmentmay clarify the principles involved.

Sources of heat in the test cell

The engine

As introduced in Chapter 2, various estimates of the total heat release to the sur-roundings from a water-cooled engine and its exhaust system have been published.One authority4 quotes a maximum of 15 per cent of the heat energy in the fuel,divided equally between convection and radiation. This would correspond to about30 per cent of the power output of a diesel engine and 40 per cent in the case ofa gasoline engine.

In the experience of the authors, a figure of 40 per cent (0.4 kW/kW engineoutput) represents a safe upper limit to be used as a basis for design for water-cooledengines. This is divided roughly in the proportion 0.1 kW/kW engine to 0.3 kW/kWexhaust system. It is thus quite sensitive to exhaust layout and insulation.

In the case of an air-cooled engine, the heat release from the engine will increaseto about 0.7 kW/kW output in the case of a diesel engine and to about 0.9 kW/kWoutput for a gasoline engine. The proportion of the heat of combustion that passes tothe cooling water in a water-cooled engine in an air-cooled unit must of course passdirectly to the surroundings.


The dynamometer

A water-cooled dynamometer, whether hydraulic or eddy current, runs at a moderatetemperature and heat losses to the cell are unlikely to exceed 5 per cent of powerinput to the brake. Usually a.c. and d.c. machines are air cooled and heat loss tosurroundings is in the region of 6–10 per cent of power input.

Other sources of heat

Effectively, all the electrical power to lights, fans and instrumentation in the test cellwill eventually appear as heat transmitted to the ventilation air. The same applies tothe power taken to drive the forced-draught ventilating fans: this is dissipated as heatin the air handled by the fans.

The secondary heat exchangers used in coolant and oil temperature control willsimilarly make a contribution to the total heat load. Here, long lengths of unlaggedpipe work holding the primary fluid will not only add heat to the ventilation loadbut will adversely effect control of temperature.

Heat losses from the cell

The temperature in an engine test cell is generally higher than usual for an industrialenvironment. There is thus in some cases appreciable transfer of heat through cellwalls and ceiling, depending on the configuration of the site but, except in the caseof a test cell forming an isolated unit, these losses may probably be neglected.

Recommended values as a basis for the design of the ventilation system are givenin Table 5.2. In all cases they refer to the maximum rated power output of the enginesto be installed.

Calculation of ventilation load

The first step is to estimate the various contributions to the heat load from engine,exhaust system, dynamometer, lights and services. This information should be sum-marized in a single flow diagram. Table 5.2 shows typical values.

Table 5.2 Heat transfer to ventilation air

kW/kW power output

Engine, water cooled 0�1Engine, air cooled 0�7–0�9Exhaust system (manifold and silencer) 0�3Hydraulic dynamometer 0�05Eddy current dynamometer 0�05a.c./d.c. dynamometer 0�15

80 Engine Testing

Heat transfer to the ventilation air is to a degree self-regulating: the cell temper-ature will rise to a level at which there is an equilibrium between heat released andheat carried away. The amount of heat carried away by a given air flow is clearlya function of the temperature rise �T from inlet to outlet.

If the total heat load is HL kW then the required air flow rate is:∗

QA =HL

1�01×1�185�T= 0�84

HL

�Tm3/s (5)

A temperature rise �T= 10�C is a reasonable basis for design. Clearly the higher thevalue of �T, the smaller the corresponding air flow. However, a reduction in air flowhas two influences on general cell temperature: the higher the outlet temperature thehigher the mean level in the cell, while a smaller air flow implies lower air velocitiesin the cell, calling for a greater temperature difference between cell surfaces and airfor a given rate of heat transfer.

Design of ventilation ducts and distribution systems

Pressure losses: fundamentals

The velocity head or pressure associated with air flowing at velocity V is given by:

pv =�V 2

2Pa

This represents the pressure necessary to generate the velocity. A typical value for�, density of air, is 1.2 kg/m3, giving:

pv = 0�6V 2 Pa

The pressure loss per metre length of a straight duct is a fairly complex function of airvelocity, duct cross-section and surface roughness. Methods of derivation with chartsare given in Ref. 5. In general, the loss lies within the range 1–10 Pa/m, the largervalues corresponding to smaller duct sizes. For test cells with individual ventilatingsystems duct lengths are usually short and these losses are small compared with thosedue to bends and fittings such as fire dampers.

The choice of duct velocity is a compromise depending on considerations of sizeof ducting, power loss and noise. If design air velocity is doubled the size of theducting is clearly reduced, but the pressure losses are increased roughly fourfold,while the noise level is greatly increased (by about 18 dB for a doubling of velocity).

Maximum duct velocities recommended in Ref. 5 are given in Table 5.3.

∗ For air at standard atmospheric conditions, �= 1.185 kg/m3. A correction may be appliedwhen density departs from this value but is probably not worthwhile.


Table 5.3 Maximum recommended duct velocities

Volume rate of flow

(m3/s)

Maximum velocity

(m/s)

Velocity pressure

(Pa)

<0.1 8–9 38–550.1–0.5 9–11 55–730.5–1.5 11–15 73–135>1.5 15–20 135–240

It is general practice to aim for a flow rate of 12m/s through noise attenuatorsbuilt within ducting.

The total pressure of an air flow pt is the sum of the velocity pressure and thestatic pressure ps (relative to atmosphere):

pt = ps+pv = ps+�V 2

2

The design process for a ventilating system includes the summation of the variouspressure losses associated with the different components and the choice of a suitablefan to develop the total pressure required to drive the air through the system.

Ducting and fittings

Various codes of practice have been produced covering the design of ventilationsystems.6�7 The following brief notes are based on the Design Notes published bythe Chartered Institution of Building Services.

Galvanized sheet steel is the most commonly used material, and ducting is readilyavailable in a range of standard sizes in rectangular or circular sections. Rectangularsection ducting has certain advantages: it can be fitted against flat surfaces and expen-sive round-to-square transition lengths for connection to components of rectangularsection such as centrifugal fan discharge flanges, filters and coolers are avoided.

Table 5.4 Maximum recommended duct velocities

Volume of flow

(m3/s)

Maximum velocity

(m/s)

Velocity pressure

(Pa)

<0.1 8–9 38–550.1–0.5 9–11 55–730.5–1.5 11–15 73–135>1.5 15–20 135–240

Pressures in ventilation systems are often measured in mmH2O:1mmH2O= 9.81 Pa.

82 Engine Testing

Once the required air flow rate has been settled and the general run of the ductingdecided, the next step is to calculate the pressure losses in the various elementsin order to specify the pressure to be developed by the fan. In most cases a cellwill require both a forced draught fan for air supply and an induced draught fanto extract the air. The two fans must be matched to maintain the cell pressure asnear as possible to atmospheric. For control purposes cell pressure is usually set at50 Pa below atmospheric which gives the safest set of conditions concerning doorpressurization and fume leakage.

Figure 5.3 shows various components in diagrammatic form and indicates the lossin total pressure associated with each. This loss is given by:

�pt = Ke

�V 2

2

d

V

V

V

V

V

hd

0.5 0.50.3 1.0

300 600 1000150

0.147

0.144

0.141

0.060

0.056

0.054

0.028

0.026 0.013

0.012

0.014

0.025

1.01.2

0.340.48

1.00.5

Entry hood

Bellmouth

Collecting hood Sudden enlargement Rectangular ductLoss = Ke

/pv

Diffuser

V

V = 20 m/s

V = 15 m/s

V = 10 m/s

V

Discharge hood Open damper

1.31.1

ha

a a1 a2

a

aa mm

10°

a1/a2

r

r

12

0.230.14

Radius bend Louvred outlet

r /ah /d

h /d

Ke

Ke Ke

Ke

Ke = 1.25

Ke = 0.15r = a /10

Ke = 0.5

Ke = 0.2

Ke = 1.25 Ke = 1.0

Ke /metre

Mitre bend

Figure 5.3 Pressure losses in components of ventilation systems


Information on pressure losses in plant items such as filters, heaters and coolers isgenerally provided by the manufacturer.

The various losses are added together to give the cumulative loss in total pressure(static pressure + velocity pressure) which determines the required fan performance.

Inlet and outlet ducting

The ducts also have to be designed so that noise created in the test cell does notbreak out into the surrounding environment. The problem has to be solved by the useof special straight attenuating sections of duct which are usually of a larger sectionand which should be designed in such a way as to give an air velocity around anoptimum value of 12m/s. Such silencer sections may tend to artificially increasethe length of ducting within the service space. Vertical intake and discharge ductingis usually more effective than horizontal wall mounted louvres in reducing noise‘break-out’ at the site border. The arrangement of air inlets and outlets calls forcareful consideration if short-circuiting and local areas of stagnant air within the cellare to be avoided.

There are many possible layouts based on combinations of high level, low level andabove engine direction ducts; four commonly used layouts are shown in Figs 5.4–5.7.

Figure 5.4 shows a commonly used design in cells without large subfloor voids;designers have to ensure that the air flow is directed over the engine rather than

Figure 5.4 Variable speed fans in a pressure balanced system using high levelinlet and outlet ducts

84 Engine Testing

Figure 5.5 Low level inlet and high level outlet ducted system; the low levelduct can either be part of a balanced fan system as in Fig. 5.4 or drawing outsideair through an inlet silencer

Figure 5.6 A balanced fan system as in Fig. 5.4, but inlet ducted throughnozzles from an over engine plenum and extracted at low level


Figure 5.7 A balanced fan system in a cell with large subfloor service spacewhere ventilation air is drawn over the engine and extracted from low level

taking a higher level short-cut between inlet and outlet. In this and the next designthe purge of floor level gases would be carried out via the exhaust gas dilution duct(see comments concerning exhaust systems in Chapter 6).

Figure 5.5 shows an alternative layout where the intake air comes into the cellat low level, either by forced ventilation, as shown, or through an attenuated ductdrawing ambient air from outside. In both cases the air is drawn over the engine andexits at high level.

An extraction inlet hood above the engine, as shown in Fig. 5.6, may inhibit theuse of service booms and other engine access.

Where a substantial subfloor space is present the ventilation air can be drawnout below floor level and over the engine from an overhead inlet ventilationplenum.

It will be clear by now that the choice of system layout has a major influence onthe layout of equipment both in the cell and the service space above.

Where an AHU is not fitted but heat energy needs to be conserved in the ventilationair then a recirculation duct, complete with flow control dampers, can be fittedbetween the outlet and inlet ducts above fans. This can be set to recirculate between0 to over 80 per cent of the total air flow.

86 Engine Testing

Fire dampers

Designs of inlet and outlet ducting should obey the rule of hazard containment;therefore fire dampers must be fitted in ducts at the cell boundary. These devices actto close off the duct and act as a fire barrier at the cell boundary.

The two types commonly used are:

• Normally closed, motorized open. These are formed by stainless steel louvreswithin a special framework having their own motorized control gearing andclosure mechanism. They are fitted with switches allowing the BMS or cellcontrol system to check their positional status.

• Normally open, thermal link retained steel curtain. These only operate when thegas temperature reaches their release temperature and have to be reset manually.

A major advantage of the motorized damper is that when the cell is not running orbeing used as a workshop space during engine rigging the flow of outside (cold) aircan be cut off.

External ducting of ventilation systems

The external termination of ducts can take various forms, often determined byarchitectural considerations, but always having to perform the function of protectingthe ducts from the ingress of rain, snow and external debris such as leaves(see diagrams of exhaust cowls in Chapter 6). To this end, the inlet often takes theform of a motorized set of louvres, interlocked with the fan start system, immediatelyin front of coarse then finer filter screens.

In some cases, in temperate climates, the whole service space above the cell maybe used as the inlet plenum for an air-handling unit; this allows more freedom inposition of the inlet louvres.

In exceptionally dusty conditions such as exist in arid desert areas the externallouvres should be motorized to close and, with vertical louvre sections, to preventdust build-up on horizontal surfaces.

The use of ‘spot fans’ for supplementary cooling

A common error made by operators facing high cell temperatures is to bring inauxiliary, high speed fans which can make the original problem worse and introducenew air flow-related problems. Spot fans can greatly increase the total amount ofheat transferred to the ventilation air, with a consequent increase in overall celltemperature. A jet or eddy of hot air can have undesirable effects: for example, itcan raise the temperature of the engine inlet air and upset the calibration of forcetransducers that are sensitive to a �t across their surface.


Control of ventilation systems

The simplest practical ventilation control system is that based on a two speed extrac-tion fan in the roof of the cell drawing air though a low level duct at one end of thecell (a variant of Fig. 5.5). The fan runs at low speed during times when the cell isused as a workshop and at high speed during engine running. In this case, the cellpressure is ambient unless the inlet filter is blocked.

Closed cells require both inlet and extract fans to be fitted with variable speeddrives and a control system to balance them. There are several alternatives, but undera commonly used control strategy operating with a non-recirculation system, one fanoperates under temperature control and the other under pressure. Thus, with bothfans running and with a cell pressure of about 50 Pa below ambient, when the cell airtemperature rises the extract fan increases in speed to increase air flow; this speedincrease creates a drop in cell pressure which is detected by the control system, sothe inlet fan speed is increased until equilibrium is restored. The roles of the fans aresomewhat interchangeable, the difference being that in the case above the controltransients tend to give a negative cell pressure and if the control roles are reversedit gives positive cell pressure.

The control fans have to be able to deal with the additional flows in and out of thecell via purge and combustion air flows where these systems are fitted. The controlsystem also has to be sufficiently damped to prevent surge and major disturbance inthe case of a cell door being opened.

Fans

Methods of testing fans and definitions of fan performance are given in BS 848. Thisis not an entirely straightforward matter; a brief summary follows.

Once again the control volume technique will be found useful. Figure 5.8 showsa fan surrounded by a control surface.

ps1

ps2

PA

QF

QFV1

V2

Figure 5.8 Centrifugal fans as an open system

88 Engine Testing

The various flows into and out of the control volume are as follows.

In air flow QF at velocity V1 and pressure ps1 power input PA

Out air flow QF at velocity V2 and pressure ps2, where ps1 and ps2 arethe static pressures at inlet and outlet, respectively.

Total pressure at inlet

pt1 = ps1+�V 2

1

2

Total pressure at outlet

pt2 = ps2+�V 2

2

2

Fan total pressure

ptF = pt2−pt1

Air power (total)

PtF =QF PtF

Most fan manufacturers quote fan static pressure which is defined as:

psF = ptF−�V 2

2

2

This ignores the velocity pressure of the air leaving the fan and does not equal thepressure difference ps2 − ps1 between the inlet and outlet static pressures. (It is worthnoting that, in the case of an axial flow fan with free inlet and outlet, fan staticpressure as defined above will be zero.)

The total air power PtF is a measure of the power required to drive the fan in theabsence of losses.

The air power (static) is given by:

PSF =Qf pSF

Manufacturers quote either fan static efficiency SA or fan total tA.The shaft power is given by:

PA =PSF

SA

=PtF

tA


Fan noise

In a ventilation system, the fan is usually the main source of system noise. The noisegenerated varies as the square of the fan pressure head so that doubling the systemresistance for a given flow rate will increase the fan sound power fourfold, or byabout 6 dB.

As a general rule, to minimize noise, ventilation fans should operate as close tothe design point as possible.

Ref. 6 gives examples of good and bad designs of fan inlet and discharge; a poordesign gives rise to increased noise generation. Ref. 8, Part 2, gives guidance onmethods of noise testing fans.

Classification of fans8

1. Axial flow fans. For a given flow rate, an axial flow fan is considerably morecompact than the corresponding centrifugal fan and fits very conveniently intoa duct of circular cross-section. The fan static pressure per stage is limited,typically to a maximum of about 600 Pa at the design point, while the fandynamic pressure is about 70 per cent of the total pressure. Fan total efficienciesare in the range 65–75 per cent. Axial flow fans mounted within a bifurcatedduct, so that the motor is external to the gas flow, are a type commonly usedin individual exhaust dilution systems. An axial flow fan is a good choice asa spot fan, or for mounting in a cell wall without ducting. Multistage units areavailable, but tend to be fairly expensive.

2. Centrifugal fans, flat blades, backward inclined. This is probably the first choicein most cases where a reasonably high pressure is required, as the constructionis cheap and efficiencies of up to 80 per cent (static) and 83 per cent (total)are attainable. A particular advantage is the immunity of the flat blade to dustcollection. Maximum pressures are in the range 1–2 kPa.

3. Centrifugal fans, backward curved. These fans are more expensive to build thanthe flat bladed type. Maximum attainable efficiencies are 2–3 per cent higher,but the fan must run faster for a given pressure and dust tends to accumulate onthe concave faces of the blades.

4. Centrifugal fans, aerofoil blades. These fans are expensive and sensitive to dust,but are capable of total efficiencies exceeding 90 per cent. There is a possibilityof discontinuities in the pressure curve due to stall at reduced flow. Theyshould be considered in the larger sizes where the savings in power cost aresignificant.

5. Centrifugal fans, forward curved blades. These fans are capable of a deliveryrate up to 2.5 times as great as that from a backward inclined fan of the samesize, but at the cost of lower efficiency, unlikely to exceed 70 per cent total. Thepower curve rises steeply if flow exceeds the design value.

The various advantages and disadvantages are summarized in Table 5.5.

90 Engine Testing

Table 5.5 Fans: advantages and disadvantages

Fan type Advantages Disadvantages

CompactConvenient installationUseful as free-standingunits

Moderate efficiencyLimited pressureFairly expensive

Axial flow

CheapCapable of highpressure

Good efficiencyInsensitive to dust

Centrifugal, flat blades,backward inclined

Higher efficiency More expensiveHigher speed for givenpressure

Sensitive to dust

Centrifugal, backwardcurved

Very high efficiency ExpensiveSensitive to dustMay stall

Centrifugal aerofoil

Small size for givenduty

Low efficiencyPossibility of overload

Centrifugal forward curved

Design of ventilation system: worked example

By way of illustration, consider the case of the 250 kW turbocharged diesel enginefor which an energy balance is given in Chapter 13, the engine to be coupled toa hydraulic dynamometer.


Ventilation air flow

Assume convection and radiation losses as follows:

engine, 10% of power output 25 kWexhaust manifold 15 kWexhaust tailpipe and silencer 15 kWdynamometer, 5% of power input 12 kWlights and services 20 kWforced draught fan 5 kW

Subtotal 92 kWLess losses from cell by conduction 5 kW

Total 87 kW

Assume air inlet temperature 20�C, �t of 11�C therefore an outlet temperature 31�C.Then air flow rate, from eq. (5), is:

QA =0�84×87

11= 6�6m3/s

+ induction air, 0.3m3/s; say 7m/s in total (= 101m3/hour per kW engine poweroutput). Assuming cell dimensions 8× 6× 4.5m high, cell volume = 216m3, thisgives 117 air changes per hour.

Table 5.2 suggests a mean duct velocity in the range 15–20m/s as appropriate, giv-ing a cross-sectional area of 0.37–0.49m2. Heinsohn2 gives a range of recommendedstandard duct dimensions of which the most suitable in the present circumstances is600× 600mm, giving a duct velocity of 19.5m/s and a velocity pressure of 228 Pa.Figure 5.9 shows one possible layout for the ventilation system.

The inlet or forced draught system uses a centrifugal fan and the duct velocityassumed above. For the extraction system, with its simpler layout and smaller pressurelosses, an axial flow fan has been chosen. The pressure losses are calculated inTable 5.6 (this process lends itself readily to computer programming) and indicatesfan duties as follows:

Forced inlet fan, flow rate 7m3/sStatic pressure 454 PaExtraction fan, flow rate 7m3/sStatic pressure 112 Pa.

A manufacturer’s catalogue offers a fan for the forced draught situation to thefollowing specification:

Centrifugal, backward inclined

Impeller diameter 900mmSpeed 850 rev/minFan static efficiency 65%

92 Engine Testing

Entryhood

Dischargehood

Centrifugalfan

2 m straight

Bend

Louvredoutlet

Collectinggrille

Axialflowfan

Sudden enlargement

Figure 5.9 Simplified cell ventilation system layout used in example calcula-tion. Purge air is taken out in combined extract duct situated low in cell

giving, from eqs (13) and (14)

shaft power=7�0×454

0�65= 4900W

motor power= 5kW

For the extraction fan, an axial flow unit has the following specification:

Diameter 1mSpeed 960 rev/minVelocity pressure 51 PaFan total efficiency 80%

shaft power=7�0× �112+51�

0�80= 1400W

In this case, the manufacturer recommends a motor rated at 2.2 kW.

Ventilation of the control room

This is in general a much less demanding exercise for the test installation designer.Heating loads are moderate, primarily associated with lights and heat generated by

Table 5.6 Calculation of system pressure losses

Item Size Area Volume

flow

rate

Velocity Velocity

pressure

Fitting

loss

factor

Pressure

drop

Length Pressure

loss

Cumulative

loss

(mm) (m2) (m3/s) (m/s) (Pa) Ke Ke/m (m) (Pa) (Pa)

Forceddraught

Entry hood 800 dia 0�5 7�0 14 118 1�1 129 129Straight 800 dia 0�5 7�0 14 118 0�02 2 5 134Bend 800 dia 0�5 7�0 14 118 0�23 27 161FanStraight 600× 600 0�36 7�0 19 217 0�025 2 11 172Bend 600× 600 0�36 7�0 19 217 0�23 50 222Suddenenlargement

600× 600 0�36 7�0 19 217 1�0 217 439

Louvred outlet 1000× 1000 1�00 7�0 7 29 0�5 15 454Total 454

ExtractionCollectinghood

1000 dia∗ 0�79 7�0 9 49 1�25 61 61

Straights 1000 dia 0�79 7�0 9 49 0�02 2 2 63FanDischargehood

1000 dia 0�79 7�0 9 49 1�0 49 112

Total 112

∗ Diameter of branch connection.

94 Engine Testing

electronic apparatus located in the room. Regulations regarding air flow rate peroccupant and general conditions of temperature and humidity are laid down in variouscodes of practice and should be equivalent to those considered appropriate for offices.

Air conditioning

Most people associate this topic with comfort levels under various conditions: sitting,office work, manual work, etc. Levels of air temperature and humidity, as well as airchange rates, are laid down by statute (see BS 57207 for particulars). Such regulationsmust be observedwith regard to the control room. However, the conditions in an enginetest cell are far removed from the normal, and justify special treatment, which follows.

Two properties of the ventilating air entering the cell (and, more particularly, ofthe induction air entering the engine) are of importance: the temperature and themoisture content. Air conditioning involves four main processes:

• heating the air;• cooling the air;• reducing the moisture content (dehumidifying);• increasing the moisture content (humidifying).

Fundamentals of psychometry

The study of the properties of moist air is known as psychometry. It is treated in manystandard texts1�2 and only a very brief summary will be given here. Air conditioningprocesses are represented on the psychometric chart (Fig. 5.10).9

This relates the following properties of moist air:

• the moisture content or specific humidity, kg moisture/kg dry air. Note thateven under fairly extreme conditions (saturated air at 30�C), the moisture contentdoes not exceed 3 per cent by weight

• the percentage saturation or relative humidity, �. This is the ratio of the mass ofwater vapour present to the mass that would be present if the air were saturatedat the same conditions of temperature and pressure. The mass of vapour undersaturated conditions is very sensitive to temperature:

Temperature �C 10 15 20 25 30Moisture content (kg/kg) 0.0076 0.0106 0.0147 0.0201 0.0273

A consequence of this relationship is the possibility of drying air by cooling.As the temperature is lowered the percentage saturation increases until at the dewpoint temperature the air is fully saturated and any further cooling results in thedeposition of moisture.

Based on a barometric

pressure of 1013 25 mbar

Spe

cific

ent

halp

y kj

/kg

Specific volume m3 / kg

Sensible/total heat

Ratio for water

Added at 30 °c

PSYCHROMETRIC

CHART

IHVE

00.1

30

0.90

90 80 70 60 50

130125120

110

105

100

95

90

85

80

75

70

60

55

50

45

40

30

20

15

10

5

0

–5

–10

25

35

65

115 135 140

Percentage saturation

40 30 20

140

13

5130

12

51

20

115

110

10

51

00

95

Mois

ture

conte

nt kg/k

g (

dry

air)

90

85

80

75

70

6560555045

403530

Specific enthalpy kj/kg

Dry – bulb temperature °c

2520

15100 5–5–10

0.030

0.029

0.028

0.027

0.026

0.025

0.024

0.023

0.022

0.021

0.020

0.019

0.018

0.017

0.016

0.015

0.014

0.013

0.012

0.011

0.010

0.009

0.008

0.007

0.006

0.005

0.004

0.003

0.002

0.001

0.00060555045403530252015100–5–10

–10

–5

0

5

10

15

20

25

0.85

0.80

0.75

5

0.2

0.3

0.4

0.5

0.6

0.70.80.91.00.90.8

0.7

0.6

0.5

0.4

0.30.2

0.1 0

Wet

– b

ulb

tem

pera

ture

°c

(slin

g)

Figure 5.10 Psychometric chart

96 Engine Testing

• The wet- and dry-bulb temperatures. The simplest method of measuring relativehumidity is by means of a wet- and dry-bulb thermometer. If unsaturated airflows past a thermometer having a wetted sleeve of cotton around the bulb thetemperature registered will be less than the actual temperature of the air, as regis-tered by the dry-bulb thermometer, owing to evaporation from the wetted sleeve.The difference between the wet-bulb and dry-bulb temperatures is a measure ofthe relative humidity. Under saturated conditions the temperatures are identical,and the depression of the wet-bulb reading increases with increasing dryness.Wet- and dry-bulb temperatures are shown in a psychometric chart.

• The specific enthalpy of the air, relative to an arbitrary zero corresponding to dry airat 0�C.We have seen earlier that on this basis the specific enthalpy of dry air

h= Cpta = 1�01ta kJ/kg (15a)

The specific enthalpy of moist air must include both the sensible heat and thelatent heat of evaporation of the moisture content.

The specific enthalpy of moist air:

h= 1�01ta +�1�86ta +2500� (15b)

the last two terms representing the sum of the sensible and latent heats of themoisture.

Taking the example of saturated air at 30�C:

h= 1�01×30+0�0273 �1�86×30+2500�

= 30�3+1�5+68�3kJ/kg

The first two terms represent the sensible heat of air plus moisture, and it isapparent that ignoring the sensible heat of the latter, as is usual in air coolingcalculations, introduces no serious error. The third term, however, representingthe latent heat of the moisture content, is much larger than the sensible heatterms. This accounts for the heavy cooling load associated with the process ofdrying air by cooling: condensation of the moisture in the air is accompanied bya massive release of latent heat.

Air conditioning processes

Heating and cooling without deposition of moisture

The expression for dry air:

H = �QACp�T kW (16)

is adequate.


Cooling to reduce moisture content

This is a very energy-intensive process and is best illustrated by a worked example.Increasing moisture content is achieved either by spraying water into the air stream(with a corresponding cooling effect) or by steam injection. With the advent ofLegionnaires’ disease the latter method, involving steam that is essentially sterile, isfavoured.

Calculation of cooling load: worked example. Consider the ventilation systemdescribed above. We have assumed an air flow rate QA = 7m

3/s, air entering at 20�C.If ambient temperature is 25�C and we are required to reduce this to 20�C then fromeq. (16) assuming saturation is not reached,

cooling load = 1�2×7�0×1�01×5= 42�4 kW

Now let us assume that the ambient air is 85 per cent saturated, �= 0.85 and thatwe need to reduce the relative humidity to 50 per cent at 20�C.

The psychometric chart shows that the initial conditions correspond to a moisturecontent:

1 = 0�0171kg/kg

The final condition, after cooling and dehumidifying, corresponds to:

2 = 0�0074kg/kg

The chart shows that the moisture content 2 corresponds to saturation at a temper-ature of 9.5�C.

From eq. (15b), the corresponding specific enthalpies are:

h1 = 1�01×25+0�0171�1�86×25+2500�

= 68�80kJ/kg

h2 = 1�01×9�5+0�0074�1�86×9�5+2500�

= 28�22kJ/kg

The corresponding cooling load:

LC = 1�2×7�0× �68�80−28�22�

= 341kW

If it is required to warm the air up to the desired inlet temperature of 20�C, thespecific enthalpy is increased to:

h3 = 1�01×20+0�0074�1�86×20+2500�

= 39kJ/kg

98 Engine Testing

The corresponding heating load:

LH = 1�2×7�0× �39�00−28�22�

= 90�5kW

To summarize:

Air flow of 7m3/s at 25�C, 85% saturatedCooling load to reduce temperature to 20�C 42.4 kWCooling load to reduce temperature to 9.5�C 341 kWwith dehumidification.Heating load to restore temperature to 20�C 90.5 kW50% saturated.

Cooling of air is usually accomplished by heat exchangers fed with chilled water.See Ref. 10 for a description of liquid chilling packages.

This illustrates the fact that any attempt to reduce humidity by cooling as opposed

to merely reducing the air temperature without reaching saturation conditions

imposes very heavy cooling loads.

Calculation of humidification load: worked example. Let us assume initial conditions,QA = 7m3/s, temperature 20�C relative humidity � = 0�3, and that we are requiredto increase this to �= 0�7 for some experimental purpose.

Relative moisture contents are:

1 = 0�0044kg/kg

2 = 0�0104kg/kg

Then rate of addition of moisture

= 7×1�2× �0�0104−0�0044�= 0�050kg/s

Taking the latent heat of steam as 2500 kJ/kg,

heat input= 2500×0�050= 125kW

This is again a very large load, calling for a boiler of at least this capacity.

Effects of humidity: a warning

Electronic equipment is extremely sensitive to moisture. Large temperature changes,when associated with high levels of humidity, can lead to the deposition of moistureon components such as circuit boards, with disastrous results.


This situation can easily arise in hot weather: the plant cools down overnight anddew is deposited on cold surfaces. Some protection may be afforded by continuousair conditioning. The use of chemical driers is possible; the granular substances usedare strongly hygroscopic11 and are capable of achieving very low relative humidities.However, their capacity for absorbing moisture is of course limited and the containershould be removed regularly for regeneration by a hot air stream.

The operational answer adopted by many facilities is to leave venerable equipmentswitched on, in a quiescent state while not in use, or to fit anti-condensation heatersat critical points in cabinets.

Legionnaires’ disease

This disease is a severe form of pneumonia and infection is usually the result ofinhaling water droplets carrying the causative bacteria (Legionella pneumophila).Factors favouring the organism in water systems are the presence of deposits such asrust, algae and sludge, a temperature between 20 and 45�C and the presence of light.Clearly all these conditions can be present in systems involving cooling towers.

Preventative measures include

• treatment of water with scale and corrosion inhibitors to prevent the build-up ofpossible nutrients for the organisms;

• use of suitable water disinfectant such as chlorine 1–2 ppm or ozone;• steam humidifiers are preferable to water spray units.

If infection is known to be present flushing, cleaning and hyperchlorination arenecessary. If a system has been out of use for some time, heating to about 70–75�Cfor 1 hour will destroy any organisms present.

This matter should be taken seriously. This is one of the few cases in which the

operators of a test facility may be held guilty of endangering life, with consequent

ruinous claims for compensation.

Combustion air treatment

The influence of the condition of the combustion air (its pressure, temperature,humidity and purity) on engine performance is discussed in Chapter 12, whereit is shown that variations in these factors can have a very substantial effect onperformance. In an ideal world engines under test would all be supplied with air at‘standard’ conditions. In practice there is a trade-off between the advantages of suchstandardization and the cost of achieving it.

For routine (non-emissions) production testing variations in the condition of theair supply are not particularly important, but the performance recorded on the testdocuments with some degree of correlation with other test beds or facilities shouldbe corrected to standard conditions.

100 Engine Testing

The simplest and most widely used method of supplying the combustion air is toallow the engine to take its supply directly from the test cell atmosphere. The greatadvantage of using cell air for vehicle engines in particular is that rigging of complex‘in-vehicle’ air filtration and ducting units is straightforward.

The major disadvantage of drawing the air from within the cell is the uncontrolledvariability in temperature and quality arising from air currents and other disturbancesin the cell. These can include contamination with exhaust and other fumes and maybe aggravated by the use of spot fans.

For research and development testing, and particularly critical exhaust emissionswork, it is necessary that, so far as is practicable, the combustion air should besupplied with the minimum of pollution and at constant conditions of temperature,pressure and humidity. While the degree of pollution is a function of the choice ofsite the other variables may be controlled in the following order of difficulty:

• temperature only;• temperature and humidity;• temperature, humidity and pressure.

Centralized combustion air supplies

Centralized combustion air conditioning units can be designed to supply air at ‘stan-dard’ conditions to a number of cells, and this may be a cost-effective solutionproviding that all cells require the same standardized conditions of combustion airand if the air is used to condition non-running cell spaces. Feedback from many sitesindicates that individual units offer the best operation solution for the majority ofresearch and development test cells.

Specification of operational envelope for humidity and temperature control

To create a realistic operational specification for temperature and humidity control ofcombustion air requires that the creator has a clear understanding of a psychometricchart (Fig. 5.10), the implications of requiring a large operational envelope plus a clearunderstanding of the temperature and humidity points actually required by the plannedtesting regimes is vital. It is recommended that cell users mark on a psychometricchart the operating points prescribed by known (emission homologation, etc.) testsor investigate the specification of available units before producing specifications thatmay impose unnecessarily high energy demands on the requested unit.

Dedicated combustion air treatment units

It should be noted that standard combustion air treatment units are sold by special-ist companies worldwide. The rating of these units matches the common sizes ofautomotive engines, but due to the different air consumption of gasoline and diesel


engines the power rating of the engines supported by one unit will differ for the twoengine types as shown in a typical example below:

Model Flow rate Suitable for engine of rating below

Unit 1 800m3/h diesel 140 kW, gasoline 200 kWUnit 2 1600m3/h diesel 280 kW, gasoline 400 kW

Temperature only control, flooded inlet

The simplest form controls air temperature only and supplies an excess of air viaa flexible duct terminating in a trumpet-shaped nozzle close to the engine’s inlet;this is known as a ‘flooded’ inlet. The unit should be designed to supply air ata constant volume, calculated by doubling the theoretical maximum air demand ofthe engine. Because the flow is constant, good temperature control can be achieved;the excess air is absorbed into the cell ventilation inlet flow and compensated for viathe ventilation temperature and pressure control.

The normal design of such units contains the following modules:

• inlet filter and fan;• chiller coils;• heater matrix;• insulated duct, via fire damper, terminating in flexible section and bellmouth.

By holding the chiller coils’ temperature at a fixed value of say 7�C and heatingthe air to the required delivery temperature, it is possible to reduce the effect of anyhumidity change during the day of testing without installing full humidity control.

Switching combustion air units on and off frequently will cause the system controlaccuracy to decline and therefore it is sensible to use the delivered air in a spaceheating role when the engine is not running allowing the main ventilation systemto be switched off or down to a low extract level. An energy-efficient comfortventilation regime can be designed by using the combustion air balanced with thepurge fan extraction.

Humidity controlled units

Systems for the supply of combustion air of which both temperature and humidityare controlled are expensive, the expense increasing with the range of conditionsto be covered and the degree of precision required. They are also energy intensive,particularly when it is necessary to reduce the humidity of the atmospheric airby cooling and condensation. Before introducing such a system, a careful analysisshould be made to ensure that the operational envelope proposed is achievable froma standard unit and the services available to it and if not, if it is really justified.

102 Engine Testing

It is sensible to consider some operational restraints on the scheduling of testshaving different operational levels of humidity in order to avoid condensation build-up in the system. Corrugated ducting should not be used in systems providing humidair since they tend to collect condensate that has to dry out before good control canbe restored.

To indicate more exactly what is involved in full conditioning of combustionair, let us suppose that it is necessary to supply air at standardized conditions tothe 250 kW diesel engine for which the energy balance is given in Chapter 13(Table 13.5).

The calculated air consumption of the engine is 1312 kg/h, corresponding to avolumetric flow rate of about 0.3m3/s.

The necessary components of an air conditioning unit for attachment to the engineair inlet are shown diagrammatically in Fig. 5.11 and may comprise in succession:

1. air inlet with screen and fan;2. heater, either electrical or hot water;3. humidifier, comprising steam or atomized water injector with associated

generator;4. cooling element, with chilled water supply;5. secondary heating element.

It may be asked why it is necessary to include two heater elements, one before andone after the humidifier. The first element is necessary when it is required to humidifycold dry air, since if steam or water spray is injected into cold air supersaturation willresult and moisture will immediately be deposited. On the other hand it is commonlynecessary, in order to dry moist air to the desired degree, to cool it to a temperaturelower than the desired final temperature and reheat must be supplied downstream ofthe cooler.

From

atmosphere

HumidifierCooler Heater

To

engine

CondensateSteam

Figure 5.11 Schematic of typical combustion air system contents


The internal dimensions of the duct in the present example would be approximately0.3× 0.3m. Assume that the air is to be supplied to the engine at the standardconditions specified in BS 5514 (ISO 3046)

temperature 25�Crelative humidity (r.h.) 30%.

Let us consider two fairly extreme atmospheric conditions and determine the condi-tioning processes necessary:

hot and humid: 35�C and 80% r.h.cool to 7�C, cooling load 34 kW(flow of condensate 0.5 l/min)reheat to 25�C, heating load 6.5 kW.

It is necessary to cool to 7�C in order to reduce the moisture content, originally0.030 kg water/kg air, to the required value, 0.006 kg water/kg air; the latter corre-sponds to saturation at 7�C:

cold and dry: 0�C and 50% r.h.heat to 25�C, heating load 9 kWinject steam at rate of 0.1 l/minheating load 4 kW.

A typical standard unit available for this type of duty has the following technicaldetails:

Unit 2 with delivery of 1600m3/h at temperature range 15–30�C with option ofhumidity control 8–20 g/kg:Total power consumption 64 kVAC.w. supply at 4–8�CPressure 2–4 barFlow rate 8m3/h.

Note that the above equates to ∼68 kW, pressure across coils approximately 120 kPa.

Demineralized water for humidifier:Temperature range 10–30�CPressure 2–4 barFlow rate 65 dm3/hCondensate drain 1/2′′

Flow rate 80 dm3/hCompressed air 6–8 barAmbient air temperature range (in order for unit to deliver air as per top line)10–35�CHumidity 3–30 g/kgSize of unit: length 2755mm × width 1050mm × height 2050mm.

104 Engine Testing

It will be observed that the energy requirements, particularly for the chiller eitherinternal or from separate chilled water supply, are quite substantial. In the case oflarge engines and gas turbines any kind of combustion air conditioning is not reallypracticable and reliance must be placed on correction factors.

Any combustion air supply system must be integrated with the fire alarm and fireextinguishing system and must include the same provisions for isolating individualcells as the main ventilation system.

Important integration points

Combustion air units of all types will require condensate drain lines with free gravitydrainage to a building system; back-up or overspill of this fluid can cause signifi-cant problems. Those fitted with steam generators will require both condensate andpossibility steam ‘blow-down’ lines piped to safe drainage.

If steam generators are fitted, the condition of feed water will be critical and

suitable water treatment plant needs to be installed, particularly where supply water

is hard.

Pressure controlled units

A close coupled air supply duct attached to the engine is necessary in special cases,such as anechoic cells, where air intake noise must be eliminated, at sites runninglegislative tests that are situated at altitudes above those where standard conditionscan be achieved and where a precisely controlled conditioned supply is needed.

Since engine air consumption varies more or less directly with speed and thatparameters can vary more rapidly than any air supply system can respond, and sinceit is essential not to impose pressure changes on the engine air, some form of excessair spill or engine bypass strategy has to be adopted to obtain pressure control withany degree of temperature and humidity stability.

For testing at steady state conditions, the pressure control system can operatewith a slow and well-damped characteristic; typically having a stabilization time,following an engine speed set-point change, of 30 seconds or more. This is wellsuited to systems based on regulating the pressure of an intake plenum fitted witha pressure controlling spill valve/flap.

The inlet duct must be of sufficient size to avoid an appreciable pressure dropduring engine acceleration and it needs to be attached to the engine inlet in sucha way as to provide a good seal without imposing forces on the engine.

It is sensible to encapsulate or connect to the normal engine air inlet filter, if itcan sustain the imposed internal pressure imposed by the supply system. If the filterhas to be encapsulated it often requires a bulky rig item that has to be suspendedfrom a frame above the engine.

Various strategies have been adopted for fast pressure control, the best of whichare able to stabilize and maintain pressure at set point within <2 seconds of anengine step change of 500 r.p.m.


Engine with

close coupled

air inlet

Fast actingpressure control

valve

Air intake atengine idle

30 kg /h1800 kg/h

1770 k

g /h

Engine silencer 1802 kg /h Extract fan

Engine air bypass

Combustion airconditioning system

TemperatureHumidity

Pressure control

Figure 5.12 AVL dynamic combustion air pressure control system. Exampleflows shown at engine idle condition

Control of inlet pressure only does not simulate true altitude; this requires thatboth inlet and exhaust are pressurized at the same pressure, a condition that can besimulated using a dynamic pressure control system patented by AVL List GmbH asshown in Fig. 5.12.

Such systems find use in motor sport and other specialist applications, but wherethe ‘ram effect’ of the vehicle’s forward motion is required, as is the case in F1 andmotor cycle engine testing, then considerable energy is required to move the air andthe test cell begins to resemble a wind tunnel, which is outside the remit of this book.

Charge (boost) air temperature control

The turbocharging process compresses the combustion air and raises its temperature,the effect of which is discussed in Chapter 12. Charge air cooling frequently has tobe accommodated in an engine test cell and there are various strategies that may beused; one is to use the ex-vehicle charge air cooler as a rig item, the other is to usea ‘fluid/charge air’ cooler as a cell fixture.

Using forced air from a spot fan to cool the vehicle charge air cooler will introducea great deal of additional heat energy into the cell; it is therefore good practice touse a liquid cooled charge air cooler having very similar flow resistance on the airside to the vehicle unit. A common strategy is to encase the original vehicle air/airintercooler in a special casing and use a water or water spray as the cooling medium.Temperature control is achieved by a flow control valve on the fluid circuit.

Summary

Design of the ventilation system for a test cell is a major undertaking and a careful andthorough analysis of expected heat loads from engine, exhaust system, dynamometer,

106 Engine Testing

instrumentation, cooling fans and lights is essential if subsequent difficulties are to beavoided. Once the maximum heat load has been determined it is recommended thata diversity factor is applied so as to arrive at a realistic rating for the cell services.

Any cell using gasoline or other volatile fuels should include a purge systemwithin the ventilation system to remove potentially explosive vapour.

The choice between full flow ventilation and air conditioning has to be made afterdetailed review of site conditions, particularly building space, and cost.

If a separate air supply for the engine is to be considered, careful thought has tobe given to its type and rating in order to keep costs and energy requirements in linewith actual test requirements.

The ventilation of the control room should ensure that conditions there meetnormal office standards.

Notation

Atmospheric pressure pa barAtmospheric temperature ta

�CDensity of air �m3

Gas constant for air R= 287 J/kgKSpecific heat of air at constant pressure Cp J/kgKMass rate of flow of air m kg/sRate of heat loss, vertical surface Qv W

2

Rate of heat loss, horizontal surface Qh W2

Temperature of surface ts�C

Rate of heat loss by radiation Qr W/m2

Ventilation air temperature rise �T �CTotal heat load HL kWVentilation air flow rate QA m

3/sFan air flow rate QF m

3/sVelocity of air Vm/sVelocity pressure pv PaStatic pressure ps PaTotal pressure pt PaPressure loss �pt PaStatic air power PsF kWTotal air power PtF kWShaft power PA kWMoisture content kg/kgRelative humidity �

Specific enthalpy of moist air h kJ/kgCooling load LC kW


Heating load LH kWEmissivity �

Pressure loss coefficient Ke

References

1. Pita, E.G. (1981) Air Conditioning Principles and Systems: An Energy Approach, Wiley,Chichester.

2. Heinsohn, R.J. (1991) Industrial Ventilation Engineering Principles, Wiley, Chichester.3. Bejan, A. (1993) Heat Transfer, Wiley, Chichester.4. Freeston, H.G. (1958) Test bed installations and engine test equipment, Proc. I. Mech. E.,

172 (7).5. TM 8 Design Notes for Ductwork, Chartered Institution of Building Services, London.6. Industrial Ventilation: A Manual of Recommended Practice (1982) (17th edn) American

Conference of Government Industrial Hygienists: Committee on Industrial Ventilation.7. BS 5720 Code of Practice for Mechanical Ventilation and Air Conditioning in Buildings.

8. BS 848 Part 1 Fans for General Purposes.

9. C.I.B.S. Psychrometric Chart, Chartered Institution of Building Services, London.10. BS 7120 Specification for Rating and Performance of Air to Liquid and Liquid to Liquid

Chilling Packages.

11. BS 2540 Specification for Granular Desiccant Silica Gel.

Further reading

BS 599 Methods of Testing Pumps.BS 6339 Specification for Dimensions of Circular Flanges for General Purpose

Industrial Fans.

6 Test cell cooling water andexhaust gas systems

Introduction

This chapter describes the systems designed to contain cooling water and exhaustgas within a test facility; the largest piped systems normally found within the enginetest cell. The thermodynamics of water cooling are considered and the calculationof cooling water requirements is described. A note on water quality follows and thedesign of test cell cooling water systems is outlined.

Options regarding design of exhaust systems are reviewed and basicrecommendations made. Finally, to pull together the various topics considered inchapters of the book relating to energy flows, an energy balance is drawn up for acomplete test cell and the process of sizing the various services is considered.

Cooling water supply system: fundamentals

The cooling water system for any heat engine test facility has provided water ofsuitable quality, temperature and pressure to allow sufficient volume to pass throughthe equipment in order to have adequate cooling capacity.

The pressure and flow rates have to be sufficiently constant to enable the devicessupplied to maintain control. A common fault of badly designed cooling watersystems is ‘cross talk’, where control of one process changes because of suddensupply pressure or temperature changes caused by external events occurring within ashared supply. It is essential for purchasers of water-cooled plant to carefully checkthe inlet water temperature specified for the required performance, since the higher thecooling water inlet temperature supplied by the factory, the less work the device willbe capable of performing before the maximum allowable exit temperature is reached.

Water

Water is the ideal liquid cooling medium. Its specific heat is higher than that of anyother liquid, roughly twice that of hydrocarbons. It is of low viscosity, relativelynon-corrosive and readily available.

Test cell cooling water and exhaust gas systems 109

The specific heat of water is usually taken as:

C = 4�1868kJ/kg K

Note: this is in fact the value of the ‘international steam table calorie’ and correspondsto the specific heat at 14�C. The specific heat of water is very slightly higher at eachend of the liquid phase range: 4.21 kJ/kg K at 0�C and at 95�C, but these variationsmay be neglected.

The use of antifreeze (ethylene glycol) as an additive to water permits operationover a wider range of coolant temperatures. A 50 per cent by volume solution ofethylene glycol in water permits operation down to a temperature of −33�C. Suchan antifreeze also raises the boiling point of the coolant and a 50 per cent solutionwill operate at a temperature of 135�C with pressurization of only 1.5 bar.

The specific heat of ethylene glycol is about 2.28 kJ/kg K and, since its densityis 1.128 kg/l, the specific heat of a 50 per cent by volume solution is:

�0�5×4�1868�+ �0�5×2�28×1�128�= 3�38kJ/kg K

or 80 per cent of that of water alone. Thus the circulation rate must be increased by25 per cent for the same heat transfer rate and temperature rise.

The relation between flow rate, qw (litres per hour), temperature rise, �t, and heattransferred to the water is:

4�1868 qw�t = 3600H

qw�t = 860H

where H= heat transfer rate, kW.(To absorb 1 kW with a temperature rise of 10�C the required flow rate is thus

86 l/h.)

Required flow rates

In the absence of a specific requirement it is good practice, for design purposes, tolimit the temperature rise of the cooling medium through the engine water jacket toabout 10�C.

In the case of the dynamometer, the flow rate is determined by the maximumpermissible cooling water outlet temperature, since it is important to avoid thedeposition of scale (temporary hardness) on the internal surfaces of the machine.Eddy current dynamometers, in which the heat to be removed is transferred throughthe loss plates, are more sensitive in this respect than hydraulic machines, in whichheat is generated directly within the cooling water.

110 Engine Testing

Recommended maximum (leaving) temperatures are:

Eddy current machines 60�CHydraulic dynamometers 70�C∗

provided carbonate hardness of water does not exceed 50mg CaO/l. For greaterhardness values limit temperatures to 50�C.

Approximate cooling loads per kilowatt of engine power output are shown inTable 6.1. Corresponding flow rates and temperature rises are as shown in Table 6.2.

Water quality1

At an early stage in planning a new test facility it is essential to ensure that asufficient supply of water of appropriate quality is made available. Control of waterquality, which includes the suppression of bacterial infections, algae and slime, is acomplex matter and it is advisable to consult a water treatment expert who is awareof local conditions. If the available water is not of suitable quality then the projectmust include the provision of water treatment plant.

Table 6.1 Cooling loads

Output (kW/kW)

Automotive gasoline engine, water jacket 0�9Automotive diesel engine, water jacket 0�7Medium speed heavy diesel engine 0�4Oil cooler 0�1Hydraulic or eddy current dynamometer 0�95

Table 6.2 Cooling water flow rates

In �C Out �C l kWh

Automotive gasoline engine 70 80 75Automotive diesel engine 70 80 60Medium speed heavy diesel engine 70 80 35Oil cooler 70 80 5Hydraulic dynamometer 20 68 20Eddy current dynamometer 20 60 20

∗ Approaching this outlet temperature, some machines experience flash boiling which can leadto a degrading of control and be heard as a distinctive ‘crackling’ noise. Prolonged runningat these conditions will cause cavitation damage to the working chamber of a dynamometer.


Most dynamometer manufacturers publish tables, prepared by a water chemist,which specify the water quality required for their machines. The following notes areintended for the guidance of non-specialists in the subject.

Solids in water

Circulating water should be as free as possible from solid impurities. If water isto be taken from a river or other natural source it should be strained and filteredbefore entering the system. Raw surface water usually has significant turbidity causedby minute clay or silt particles which are ionized and may only be removed byspecialized treatments (coagulation and flocculation). Other sources of impuritiesinclude drainage of dirty surface water into the sump, windblown sand enteringcooling towers and casting sand from engine water jackets. Hydraulic dynamome-ters are sensitive to abrasive particles and accepted figures for the permissiblelevel of suspended solids are in the range 2–5mg/l. The use of seawater or estu-arine water is not to be recommended other than in specially designed marineinstallations.

Water hardness

The hardness of water is a complex property, while there is a general subjectiveunderstanding of the term related to the ease with which soap lather can be createdin a water sample; the quality is not easy to measure objectively. Hard water, ifits temperature exceeds about 70�C, may deposit calcium carbide ‘scale’, whichcan be very destructive to all types of dynamometer and heat exchanger. A scaledeposit greatly interferes with heat transfer and commonly breaks off into the waterflow when it can jam control valves and block passages. Soft water may havecharacteristics that cause corrosion, so very soft water is not ideal either.

Essentially, hardness is due to the presence of divalent cations, usually calciumor magnesium, in the water. When a sample of water contains more than 120mg ofthese ions per litre, expressed in terms of calcium carbonate, CaCO3, it is generallyclassified as a hard water.

There are several national scales for expressing hardness, but at present nointernationally agreed scale:

American and British: 1� US= 1� UK= 1mg CaCO3 per kg water= 1 ppm CaCO3

French: 1� F= 10mg CaCO3 per litre waterGerman: 1� G= 10mg CaCO3 per litre water1� dH= 10mg CaO per litre water= 1.25� English hardness(the old British system, 1 Clarke degree= 1 grain per Imperial gallon= 14.25 ppm

CaCO3).

Requirements for dynamometers are usually specified as within the range 2–5 Clarkedegrees (30 to 70 ppm CaCO3).

112 Engine Testing

Water may be either acid or alkaline/basic. Water molecules, HOH (commonlywritten H2O), have the ability to dissociate, or ionize, very slightly. In a perfectlyneutral water (neither acidic nor basic) equal concentrations of H+ and OH− arepresent. The pH value is a measure of the hydrogen ion concentration: its value isimportant in almost all phases of water treatment, including biological treatments.Acid water has a pH value of less than 7.07 and most dynamometer manufacturerscall for a pH value in the range 7–9; the ideal is within the range 8–8.4.

The preparation of a full specification of the chemical and biological properties ofa given water supply is a complex matter. Many compounds – phosphates, sulphates,sodium chloride and carbonic anhydride – all contribute to the nature of the water;the anhydrides in particular being a source of dissolved oxygen that may make itaggressively corrosive. This can lead to such problems as the severe roughening of theloss plate passages in eddy current dynamometers, causing failures due to local waterstarvation and plate distortion (the narrow passages in eddy current dynamometersare particularly liable to blockage which can arise from chemicals used in somewater treatment regimes). Water treatment specifications should include the fact that,if used with water brakes, the treated water will be subjected to highly centrifugalregimes and local heating that may cause some degrading of the solution.

Control of water quality also includes the suppression of bacterial infections, algaeand slime. BS 49592 describes the additives used to prevent corrosion and scaleformation, with chemical tests for the control of their concentration and also givesguidance on the maintenance and cleaning of cooling water systems.

A recirculating system should include a small bleed-off to drain, to preventdeterioration of the water by concentration of undesirable elements. A bleed rate ofabout 1 per cent of system capacity per day should be adequate. If no bleed-off isincluded, the entire system should be periodically drained, cleaned out and refilledwith fresh water.

Finally, consideration should be given at the design stage to the consequencesof a power failure. Consider, for example, the consequences of a sudden failure inthe water supply to a hydraulic dynamometer absorbing 10MW at full speed from amarine diesel engine. The system will take some time to bring to rest, during whichthe brake will be operating on a mixture of air and water vapour, with the possibilityof serious overheating. In the case of large engine test facilities some provision fora gravity feed of water in the event of a sudden power failure is advisable.

It is desirable that the supply pressure to hydraulic dynamometers should bestable or the control of the machine will be affected. This implies that the supplypump must be of adequate capacity, with a flat pressure–volume characteristic.Sufficient pressure must be available under maximum demand conditions to meetthe specified dynamometer head. The design and installation of a closed (see point3 below) water supply for a large test installation is a specialist task not to beunderestimated. It may require the inclusion of a large number of test and flowbalancing valves, together with air bleed points, stand-by pumps and filters withchangeover arrangements.


Types of test cell cooling water circuits

Engine coolant control systems and cell cooling water circuits may be classified asfollows, with increasing levels of complexity:

1. Direct mains water supplied systems containing service modules and coolingcolumns that allow heated water to run to waste.

2. Sump or tank stored water systems that are ‘open’, meaning at some point in thecircuit water runs back into the sump via an open pipe. These systems normallyincorporate self-regulating water/fluid cooling modules for closed engine coolingsystems filled with special coolant/water mix and, if required, for lubricantcooling. They commonly have secondary pumps to circulate water from thesump through evaporative cooling towers when required.

3. Closed pumped circuits with an expansion, pressurization and make-up unitsin the circuit. Such systems have become the most common as most moderntemperature control devices and eddy-current or electrical dynamometers, unlikewater-brakes, do not require gravitational discharge. Closed water cooling sys-tems are less prone to environmental problems such as Legionnaires’ disease.

4. Chilled water systems (those supplying water below ambient) are almost alwaysclosed.

‘Open’ water cooling circuits

The essential features of these systems are that they store water in a sump lying belowfloor level from which it is pumped through the various heat exchangers and a coolingtower circulation system. The sump is normally divided into hot and cold areas by apartition weir wall (see Fig. 6.1). Water is circulated from the cool side and drainsback into the hot side. When the system temperature reached the control maximum,it is pumped through the cooling tower before draining back into the cool side.A rough rule for deciding sump capacity is that the water should not be turned overmore than once per minute. Within the restraints of cost, the largest available volumegives the best results. Sufficient excess sump capacity, above working level, should

be provided to accommodate drain-back from pipework, engines and dynamometers

upon system shutdown.

There is a continuous loss of water due to evaporation plus the small drainage towaste mentioned above and make-up is supplied by way of a float valve fitted to amains water supply.

To minimize air entrainment the pump suction should be located close to a corner;return flow should be by way of a submerged pipe with air vent.

This is a classic arrangement with thousands of similar systems installedworldwide, but care has to be taken to keep debris such as leaves or flood water‘wash-off’ from entering the system via the sump lip or the cooling tower collector.A sensible design feature at sites where freezing conditions are experienced is to use

114 Engine Testing

Engine

Cooling

tower

Cooling tower

pumpCooling water

circulation pumpEngine coolant

temperature control unit

Make up supply

Hot sump

Cool

sump

Figure 6.1 Simple open cooling water system incorporating a partitioned sump

pumps submerged in the sump so it can be ensured that, when not being used, themajority of pipe work will be empty.

Closed water cooling circuits

Essentially, the system uses one or more pumps to force water through the circuitload where it picks up heat which is then dispersed, usually in the air, via closed‘cooling towers’, then the water is returned directly to the pump inlet.

It is vital that air is taken out and kept out of the system and that the whole pipesystem be provided with the means of bleeding air out at high points or any trappoints in the circuit. To achieve proper circulation, to cope with thermally inducedchanges of system volume and to make up for any leakage, the closed system has tobe fitted with an expansion tank and means of pressurization. Both these requirementscan be met by using a form of compressed air/water accumulator connected to a‘make-up’ supply of treated water. Balancing water systems is the means by whichthe required flow, through discrete parts of the circuit having their own particularresistance to flow, is fixed by use of flow regulation valves having test points fittedfor setting purposes. The balancing of closed cooling systems can be problematic,particularly if a facility is being brought into commission in several phases meaningthat the complete system will have to be rebalanced at each significant change.


None of the devices fitted within a closed and balanced plant water system shouldhave valves that change the flow of the plant water (economizer valves) since thatvariation will continually unbalance the whole system.

Closed systems are often filled with an ethylene glycol–water mix to cope withfreezing weather conditions or have any external pipe work trace heated.

Cooling columns

If special engine coolants are not required, a cooling column is a simple andeconomical solution commonly used in the USA. It can be portable and locatedclose to the engine under test. This allows the engine outlet temperature to run up toits designed level; at this operating point a thermostatic valve opens, allowing coldwater to enter the bottom of the column and hot water to run to waste or the sumpfrom the top. The top of the column is fitted with a standard automotive radiator capfor correct pressurization and use when filling the engine circuit.

Engine coolant and oil temperature control modules

Whether or not the engine under test is fitted with its own thermostat, precise controlof coolant temperature is not easily achieved unless the service module used isdesigned to match the thermal characteristics of the engine with which it is associated;even here it may be difficult to achieve stable temperatures at light load.

The instability of temperature control is increased if the engine is much smallerthan that for which the cooling circuit is designed. The capacity of the heat exchangeris the governing factor and it may be advisable, when a wide range of engine powersis to be accommodated, to provide several coolers with a range of capacities.

There are many closed system engine coolant temperature control units on themarket, most working on the principle of a closed loop control valve controlling flowof coolant through a heat exchanger and they can be broken down into the followingtypes:

1. mobile pedestal type;2. special engine pallet-mounted systems;3. user-specific, wall-mounted systems;4. complex fixed pedestal type.

Figure 6.2 is an illustration of a typical service module incorporating heat exchangersfor jacket coolant and lubricating oil, while Fig. 6.3 shows a simplified schematic ofthe circuit.

The combined header tank and heat exchanger is a particularly useful feature.This has a filler cap and relief valve and acts in every way as the equivalent ofa conventional engine radiator, ensuring that the correct pressure is maintained.If some engines are to be tested without their own coolant pumps the module must

116 Engine Testing

Air release valve

Combined header tank

and heat exchanger

3 way control valve

Engine oil

sump level

Oil temperature

control valve

Heat exchanger

Plant water

supply

and return

Flow setting

valve in

plant water

circuit

OUT

+

+

+

+

+

+

+ +

+ +

+

+

+ +

+

+

+

+

+

+

+ +

+ +

+

+

++

++

+

Figure 6.2 Diagram of movable coolant conditioning unit with combined heatexchanger/header tank

be fitted with a circulating pump, commonly of the type used in central heatingsystems. For ease of maintenance, it should be possible to withdraw exchanger tubestacks without major dismantling of the system and a simple means for draining bothoil and a coolant circuits should be provided.


Enginecoolant in

Heat exchangerheader tank

Oilcooler

Plantwater

Enginecoolant return

Engineoil in

Engineoil return

Figure 6.3 Circuit diagram of unit in Fig. 6.2

The most usual arrangement is to control the temperature by means of a three-waythermostatically controlled valve in the engine fluid system. The alternative, wheretemperature is controlled by regulating the primary cooling water flow, will workbut gives an inherently lower rate of response to load changes.

The types 2 and 3 listed above are often designed and built by the user, particularlythe pallet-mounted systems which may use specific ex-vehicle parts for such itemsas the header tank and expansion vessel.

Type 4 are the most complex and incorporate a coolant circulation pump, heatersand complex control strategies to deal with low engine loads and transient testing.

None of these devices will operate satisfactorily if not integrated well with theengine and cell pipe work.

Time and distance lag, between a sensor located at the engine (inlet or outlet) andthe control valve at the cooler, may be significant and the length and volume of piperuns between engine and service module should be kept to a minimum. The sum ofthese phenomena is often referred to as the ‘thermal inertia’ of the cooling systemand can be most easily visualized by considering the speed at which heated or cooledfluid is circulated, detected and diverted within the total engine/cooling system.

118 Engine Testing

To reduce thermal inertia there are two widely used strategies:

1. Reduce the distance and fluid friction head between coolers and engine.2. Circulate the coolant between engine and coolers with auxiliary pumping (in

both cases the interconnecting pipes should be insulated against heat loss/gain).

Strategy 1 is best served by arranging a pallet-mounted cooling module close to theengine.

In cases such as anechoic cells, where the heat exchanger is inevitably remotefrom the engine, strategy 2 is required to speed up the rate of circulation by anauxiliary pump mounted outside the cell to reduce lag.

Some high-end devices, such as the AVL 553®, provides a continuous circulationof coolant within its own system from which the engine draws off the required flow.However, the quality of control is still dependent on good installation to reduce fluidtransit time and heat loss.

Control strategies of heat exchanger systems

Three strategies are commonly adopted for the source of control:

• independent packaged controllers using electrically operated valves such thoseproduced by Honeywell;

• indirect control by instruments having internal control ‘intelligence’ and givenset points from a central (engine test) control system;

• direct control from the engine test controller via auxiliary PID software routinesand pneumatically operated valves.

A proportional and integral (P and I) controller will give satisfactory results in mostcases, but it is important that thermostatic valves should be correctly sized, since theywill not function satisfactorily at flow rates much lower than that for which they aredesigned. In design terms, the valves have to have ‘authority’ over the range of flowsthat the coolers require. A common error, which creates a circuit where valves do nothave authority at low demands, is to fit valves that have too higher rated flow rate; itis very difficult to control such a system at low heat levels, since the valve only hasto crack open for an excess of cooling to take place and stability is never achieved.

Ensure that the valve is installed with flow in the correct direction!

Flow velocities in cooling water systems

The maximum water velocity in a supply system should not exceed 3m/s, but mayneed to reach a minimum of 1.5m/s in some systems to sweep away deposited matter.Velocities in gravity drains will not normally exceed 0.6 m/s. It is good practice tokeep pipe runs as straight as possible and to use bends rather than elbows.


Design of water-to-water heat exchangers

Manufacturers of heat exchangers and devices using cooling water invariably providesimple design procedures for establishing the sizes of flow rates and pressure dropfor a given heat exchanger performance. If it should be required to design a heatexchanger from first principles, Ref. 3 gives a detailed design procedure.

Engine thermal shock testing

To accelerate durability testing of engines and engine components, including cylinderhead seals, many manufacturers carry out thermal shock tests which may take severalforms, but commonly require the sudden exchange of hot coolant within a runningengine for cold fluid. The term ‘deep thermal shock’ is usually reserved for testshaving a �t of around 100�C; such tests are also called, descriptively, ‘head cracking’tests. All such sudden changes in fluid temperatures cause significant differentialmovement between seal faces.

Thermal shock tests are normally carried out in normal (not climatic) test cellsand are achieved by having a source of cold fluids that can be switched directlyinto the engine cooling circuit via three-way valves inserted into the coolant systeminlet and outlet pipes. Due to the speed required to chill down the engine and thecyclic nature of the tests it is not usual to use secondary heat exchanges between thechilled fluid and the engine coolant. The energy requirements for such tests clearlydepend on the size of engine and the cycle time between hot stabilized running andthe chilled engine condition, but they can be very considerable and often requirecells with specially adapted fluid services that include a large buffer tank of chilledcoolant. Mobile thermal shock systems, complete with chiller and buffer tank, aremade and allow any cell fitted with suitable actuating valves in the engine coolantcircuit to be used for thermal shock testing. However, their capacity is less than thatof a purposely adapted cell.

Commissioning of cooling water circuits

Before dynamometers and any other test cell instrumentation are connected tothe building’s cooling water system, it must be cleaned and the water treatedappropriately. Good practice dictates that during water system commissioning thesystem is first fully flushed. At each point that is as close as possible to an instrumentthat has to be supplied with water, the system should be fitted with temporarybypass pipes that allow water to be circulated. If this is not possible or if large heatexchangers have to be flushed then temporary strainers should be put in circuit.

Only when temporary strainers and filters are not picking up debris shouldinstruments be connected.

In locations where water is freely available it is recommended that the first systemfill of untreated mains supplied water used for flushing is dumped and the settlingtanks cleaned if necessary. Debris will include both magnetic and non-magnetic

120 Engine Testing

material injurious to test plant, such as jointing compound, PTFE tape, weldingslag and building dust. It is cost-effective to take trouble in getting the coolant upto specification at this stage rather than dealing with the consequences of valvemalfunction or medium-term corrosion/erosion problems later.

After flushing, the system should then be filled with clean water that hasbeen treated to balance the hardness, pH level and biocide level to the requiredspecification.

Drawing up the energy balance and sizing the system

In Chapter 2 it is recommended that at an early stage in the design of a new test cella diagram similar to Fig. 6.4 should be drawn up to show all the flows: air, fuel,water, exhaust gas, electricity and heat into and out of the cell.

The process is best illustrated by an example, and the full-load regime for a 250 kWturbocharged diesel engine driving a hydraulic dynamometer has been chosen. Theheat balance for the engine itself is shown diagrammatically in Chapter 13. Thisshows the results of the engine cooling water calculation, the exhaust energy, theinduction air flow and the estimated convection and radiation losses.

The design of the ventilation system for the cell is dealt with in Chapter 5, andthe corresponding air and heat flows are shown in Chapter 13.

The cooling water flow for the hydraulic dynamometer may be calculated on thebasis of the following assumptions:

• 95 per cent of power absorbed appears in the cooling water;• cooling water inlet temperature 30�C;• maximum desirable outlet temperature 70�C.

Then flow to brake:

250×0�95×60

4�186×40= 851 l/min

The electrical power input to the cell for lights, instrumentation and spot fans isassumed to be 20 kW and, finally, the heat losses through the cell walls and ceilingare assumed to be small, 5 kW.

Calculated pipe sizes are as follows:

Fuel, velocity 0.2m/s, 10.8mm dia, say 0.5 inEngine cooling water, 3m/s, 45mm dia, say 2 inBrake cooling water, 3m/s, 28mm dia, say 1.5 in

A final point regarding engine cooling water: it may be assumed that this will makeuse of a service module. Either way, the overall energy flow will be unchanged butin the case where one is not used the temperature rise of the primary cooling watermay be substantially higher, leading to a smaller flow rate.

Forceddraughtfan

Fuel592 kW58 I/h

Induction air0.3 m3/s

13 mm(½ in)

50 mm(2 in)

Enginecoolingwater

In

In

Turbochargeddiesel engine

40 kW

15 kW

12 kW

15 kW

Contr

ol

console

Hydraulicdynamometer Electrical

services20 kW

Cell walllosses5 kW

Dynamometercontrol surface

Lights5 kW

Exhaust177 kW

Power 250 kW

25 mm (1 in)

113 I/min30°C

+238 kW60°C

Enginecontrolsurface

Brakecoolingwater

Exhaust162 kW Cell control

surface

Ventilationair out+87 kW5.9 m3/s38°C

Ventilationair in+5 kW6.2 m3/s25°C

Out

Out

+ 125 kW80°C

290 I/min70°C

Figure 6.4 Energy balance and energy flow diagram for example 250 kW cell

122 Engine Testing

Exhaust gas systems

The layout and design of exhaust systems forming part of full gaseous emissionanalysis systems is covered in Chapter 16.

It is possible to run into operational safety problems with test cell exhaust systems;there have been accidents because of these systems, some of them fatal.

Particular care should be taken in the detailed design and construction of exhaustducting containing exhaust gases under positive pressure within building spaces.Within the test cell the most significant risk to operation staff is contact with hot metalexhaust piping. From the research engineer’s viewpoint, the ideal exhaust system foran engine under test is one that resembles exactly the system that would be usedon the same engine in service; indeed, many modern automotive engines are fittedwith emission control systems which require that the actual vehicle exhaust systemshould be employed. This can impose problems of layout even in cells of largefloor area. Similarly, the practical problems of locating the tail pipe in the scavengeduct can be considerable and often calls for modification of the vehicle system orflexible extension to the cell system. If vehicle systems are to be modified, it shouldbe remembered that any change in the length of the primary pipe is particularlyundesirable, since this can lead to changes in the exhaust and induction processesas a result of changes in the pattern of exhaust pulses in the system. This canaffect the volumetric efficiency and power output of the engine and, in the case oftwo-stroke engines, it may prove impossible to run the engine at all with a wronglyproportioned exhaust system. An increase in the length of pipe beyond the firstchamber is less critical, but care should be taken to limit the back pressure imposedupon the engine, as measured by a manometer at the silencer inlet, to a maximum ofperhaps 100mmH2O; any greater restriction will probably have a perceptible effecton volumetric efficiency and power output.

Turbocharged engines in particular have complex exhaust systems and run atsuch high temperatures that large areas of manifold and exhaust pipe can appearincandescent and this can represent a large heat load on the ventilating system. Theparts of the exhaust pipe system after the engine, particularly running at low level inthe cell should be guarded to avoid burns to staff and vulnerable equipment. The useof bandage-type lagging may be appropriate providing it does not cause overheatingof exhaust system components. Unless the system has been specifically designed torun the exhaust pipe in a floor duct and housekeeping is of a high order, it is notgenerally a good idea.

The practice sometimes adopted of discharging the engine exhaust into the mainventilation extraction duct has several disadvantages:

• The duct should be of stainless steel, to avoid rapid corrosion.• The air flow must be increased to a level greater than that necessary for basic

ventilation, to maintain an acceptable duct temperature. On a cold day this canlead to chilling in the cell.


• Soot deposits and staining condensate in the fan and ducting are unsightly andmake maintenance difficult.

• Other difficulties can arise, such as noise, variability in exhaust back pressure,etc.

If ‘tail pipe’ testers of the type used in vehicle servicing and inspection are to beemployed, then easy access to the tail pipe end is necessary.

Test cell exhaust layouts may be classified as follows:

• individual cell, close coupled;• individual cell, vehicle exhaust system, scavenged duct;• multiple cells with common scavenged duct;• specially designed emission cells (see Chapter 16).

Individual cell, close coupled

Such an arrangement is shown schematically in Fig. 6.5a, it may be regarded as the‘standard’ arrangement for a general purpose test bed, and is also commonly usedfor production testing. The exhaust manifold is coupled to a flexible stainless steelpipe, of fairly large diameter to minimize pressure waves, and led by way of a backpressure regulating valve to a pipe system suspended from the cell roof. Condensate,which is highly corrosive, tends to collect in these pipes, which should be laid to afall with suitable drainage arrangements.

In cells designed to occasionally use smoke or particulate analysis, it is usuallya requirement that the sample is taken from the exhaust pipe at particular points;some devices have quite specific requirements regarding size, position and angle ofprobe insertion and it is desirable, in order to ensure a representative sample, that

(a) Individual cell, close coupled

Z

(b) Individual cell, engine system and scavenge duct

Fan

(c) Multiple cells with common scavenge duct

Figure 6.5 Exhaust systems

124 Engine Testing

there should be six diameters length of straight pipe both upstream and downstreamof the probe.

It is therefore good practice, when designing the exhaust system, to arrange astraight horizontal run of exhaust pipe; this pipe to be easily replaced by a pipe withspecific probe tappings.

Individual cell, engine system scavenged duct

When it is considered necessary to use the vehicle exhaust system, two options areavailable: one is to take the pipe outside the building through a panel in the cellwall, the other is to use a scavenge air system as shown schematically in Fig. 6.5b.In this case, the tail pipe is simply inserted into a bellmouth through which cell air isdrawn; the flow rate should be at least twice the maximum exhaust flow, preferablymore. This outflow should be included in the calculations of cell ventilation airflow. The scavenge flow is induced by a fan, usually centrifugal, which must becapable of handling the combined air and exhaust flow at temperatures that mayreach 150�C.

Multiple cells common scavenged duct

This arrangement, Fig. 6.5c, may be regarded as standard for large installations.To illustrate the design process for the scavenging system, let us suppose we aresetting up an installation of three test cells each like that schematically illustratedin Fig. 6.4. These are to be used for the quality audit testing of the 250 kWturbocharged diesel engine for which maximum exhaust flow rate is shown in theenergy balance calculation (see Chapter 13) as 1365 kg/h. To ensure adequate dilutionand a sufficiently low temperature in all circumstances, we need to cater for thepossibility of all three engines running at full power simultaneously and a scavengeair flow rate of about 10 000 kg/h, say 2.3m3/s would be appropriate. Table 5.3(Chapter 5) indicates a flow velocity in the range 15–20m/s and hence a duct sizein the region of 400× 300mm, or 400mm diameter. As before, the scavenging fanmust be suitable for temperatures of at least 150�C.

This arrangement is recommended only for diesel engines. In the case of sparkignition engines there is always the possibility that unburned fuel, say from an enginethat is being motored, could accumulate in the ducting and then be ignited by theexhaust from another engine. The possibility may seem remote, but accidents of thiskind are by no means unknown.

Note particularly that, in cases (b) and (c), the fan controls must be spark-proofand interlocked with the cell control systems so that engines can only be run whenthe duct is being evacuated.

Where the engine’s exhaust system is not used, the section of exhaust tubingadjacent to the engine must be flexible enough to allow the engine to move onits mountings and a stainless steel bellows section is to be recommended. Exhaust


tubing used in this area should be regarded as expendable and the workshop shouldbe equipped to make up replacements and fit transducers.

As a final point, carbon steel silencers that are much oversized for the capacityof the engine will never get really hot and can be rapidly destroyed by corrosion.

Dual use of exhaust gas extraction duct as cell purge systems

Exhaust gas dilution duct fans compliant with European ATEX rules (see Chapter 4)can be integrated into a system design so as to be used as low level hydrocarbonextract/purge systems that run before engine operation, or at a high hydrocarbon gaslevel alarm. Figure 6.6a shows a possible design for subfloor, close coupled exhaustsystems and (b) shows a possible layout for connection to an, above floor, vehicleexhaust system.

Cooling of exhaust gases

There may be an operational requirement to reduce the temperature of exhaust gasexiting the cell. This is usually achieved by using a stainless steel, water-jacketedcooler having several gas tubes that give a low resistance to gas flow. Theseoccasionally used devices are often made specifically for the project in which theyare used and there are three important design considerations to keep in mind:

• If water flow is cut off to the exhaust gas cooler, a steam explosion can be caused,therefore primary and secondary safety devices should be fitted.

• High internal stresses can be generated by the differential heating of the coolerelements therefore some expansion of the outer casing should be built into thedesign.

• Given water jacket pressure will be higher than exhaust gas pressure, tube leakagecould fill the engine cylinders unless the system is regularly checked.

Direct water spray has been used in vertical sections of exhaust systems in order towash and cool exhaust flow into a common evacuated exhaust main.

Exhaust cowls on buildings

When deciding on the position of the exhaust termination outside a building, it isimportant to consider the possibility of recirculation of exhaust fumes into ventilationinlets and also to avoid the imposition of back pressure through wind flow. Preventionof recirculation requires the careful relative positioning of exhausts and inlet ducts inrelation to each other and to the prevailing local (building effected) wind direction.

Figure 6.7a shows a simple termination tube suitable for use on single cell, rawexhaust outlets that incorporates a shroud tube that acts as a plume dilution device and

126 Engine Testing

(a)

(b)

Figure 6.6 (a) Layout of subfloor exhaust collection duct suitable for left- orright-handed or Vee exhaust configurations and tapping for purge system ofsubfloor space. (b) Layout of above-floor exhaust dilution duct for collection tovehicle exhaust system and low level purge system


Raindrain

(b)(a)

Figure 6.7 (a) A simple rain excluding dilution tube fitted to a single cell rawexhaust outlet and (b) one design of a building cowl for the termination of amulticell diluted exhaust system

rain excluder. Figure 6.7b shows a type of rain excluding cowl commonly used onmultiple cell diluted exhaust outlets. In both cases, the exit tubes terminate verticallyto minimize horizontal noise spread.

Summary

The design of various services concerned with cell cooling water and engine exhaustis outlined. Precautions regarding water are discussed and finally an example ofcooling water system design and sizing is given. The possible dangers associatedwith exhaust systems are emphasized and a number of layouts discussed.

128 Engine Testing

References

1. Vesilind, P.A. Weiner, R.F. and Peirce, J.J. (1988) Environmental Engineering,Butterworth-Heinemann, Oxford.

2. BS4959 Recommendations for Corrosion and Scale Prevention in Engine Cooling Systems.

3. McAdams, W.H. (1973) Heat Transmission, McGraw-Hill, Maidenhead.

7 Fuel and oil storage, supplyand treatment

Introduction

Liquid fuels must be stored and transported in conditions which minimize theloss of volatile components, an effect known as ‘fuel aging’. The storage andtransport of volatile liquid and gaseous fuels worldwide is subject to extensiveand ever-developing regulation covering both safety and environmental issues. Whilemuch of the legislation quoted in this chapter is British or European in origin, thepractices they require are valid in most countries in the world.

In the UK, the Health and Safety Commission (HSC) and the Health and SafetyExecutive (HSE) are responsible for the regulation of almost all the risks to healthand safety arising from work activity, while 410 local authorities (LA) in England,Scotland and Wales have responsibility for the enforcement of health and safetylegislation.

The HSE introduced the Dangerous Substances and Explosive AtmospheresRegulations in 2002. Approved Codes of Practice supporting this new legislationrequire underground storage tanks to be provided with secondary containment ora leak detection system capable of identifying leaks before a hazardous situationcan arise. There are European regulations such as the ATEX Directive-94/9/EC, butindividual countries within the EEC will then have a level of local regulation, as isthe case in different administrative areas in the USA.

Bulk fuel storage and supply systems

In England, if fuel (such as petrol, diesel, vegetable, synthetic or mineral oil) isstored in a container with a storage capacity of over 200 litres (44 gallons) then theowner may need to comply with the Control of Pollution (Oil Storage) (England)Regulations 2001. Similar regulations are in place throughout Europe. Many of thelocal authorities act as a local Petroleum Licensing Authority. To gain a licence, thefollowing information will usually be required:

1. Site location map to a scale of 1:1250 or 1:2500 indicating all site boundaries.2. Two copies of site layout to scale 1:100 clearly indicating the present layout of

petrol storage depot.

130 Engine Testing

The layout plan must show the following:

• location of storage tanks and tank capacities;• route taken by road delivery vehicles (this must be a ‘drive-through’ not a

‘cul-de-sac’);• position of fill points and their identification;• location of pipework including all vent pipes, etc.;• location of metering pumps, dispensers, etc.;• all site drainage and its discharge location;• petrol interceptor location and drainage discharge point;• all other buildings within the site and their use;• all neighbouring buildings within 6m from the boundary;• position of LPG storage, if applicable;• location of car wash and drainage if applicable;• main electrical intake point and distribution board;• position of all fire fighting appliances.

Very similar restrictions and requirements to those described above for the UK existin most countries in the world and any engineers taking responsibility, for anythingbeyond the smallest and simplest test cell, should make themselves familiar withall relevant regulations of this kind, both national and local. Any contractor used toinstall or modify fuel systems should be selected on the basis of proven competenceand, where applicable, relevant licensing to carry out such work.

Both the legislative requirements and the way in which they are interpreted by localofficials can vary widely, even in a single country. Where engine test installations arewell established there should be a good understanding of the requirements, whereasin localities where such systems are novel, the fire and planning officers may haveno experience of the industry and can react with concern; in these cases they mayrequire tactful guidance.

The author refers readers back to the statement made in Chapter 1 concerning

the cell being a hazard reduction and containment box.

The vast majority of engine test cells are designed to be classified as zone 2 spaces.The interior of instrument cabinets such as fuel consumption or fuel conditioningdevices should be certified by the manufacturer as zone 2.

Figure 7.1 shows a typical arrangement for a fuel oil or gasoline bunded storagetank in accordance with British Standard 799.

The risk of oil being lost during filling or from ancillary pipework is higher thantank rupture; the UK Control of Pollution (Oil Storage) Regulations 2001 recognizethis fact and require that tanks have all ancillary equipment such as sight tubes, tapsand valves retained within the secondary containment system.

The use of double-skinned tanks is strongly recommended when used inunderground installations, as is common in the case for bulk storage of petroleum.

Fuel and oil storage, supply and treatment 131

Oil return line forunloading pumpedring main systems

Alternative oilsubmersionheater

Thermostat foralternative heater Vent pipe

Fillingline

Fillingpoint

Heater elementwithdrawal pointOil draw-offline

Fire valve

Filter

Outletvalve

Oilimmersion

heater

Oil tightlining

Tank supportsCatchpitGrating

Semi-rotary pump

Tank laggingif required

Fire valve wireManhole coverwith provisionfor dipstick

Figure 7.1 Typical above ground bunded fuel tank with fittings

Such tanks should also be fitted with an interstitial monitoring device withautomatic alarms.

For the storage of fuel in drums or other containers, a petroleum store such as isshown in Fig. 7.2 is to be recommended. This design typically meets the requirementsof the local UK authorities, but its location and the volume of fuel allowed to bestored may be subject to site-specific regulation. The lower part forms a bund, notmore than 0.6m deep, capable of containing the total volume of fuel authorized tobe kept in the store. Ventilators at high and low level are to be covered by finewire gauze mesh and protective grilles. Proprietary designs should allow access toa wheeled drum lifter/trolley.

127 mm (5 in) concrete

VentVent

Fire-resisting door

229 mm (9 in) brickwork

Figure 7.2 Design of fuel drum store minus statutory warning signs

132 Engine Testing

Fuel pipes

The expensive consequences of environmental pollution caused by fuel leaks aretending to cause modern best practice to dictate that fuel lines are run above ground,where they can be seen, rather than in underground trenches which was the commonpractice before the 1990s.

The use of standard (non-galvanized) drawn steel tubing for fuel lines that remainfull of liquid fuel is entirely satisfactory, but if they are likely to spend appreciableperiods partially drained the use of stainless steel is recommended. The use ofthreaded fittings is not to be recommended, although provided there is no significantpipe movement, particularly thermally induced movement, and with the use of modernsealants they can be satisfactory. The use of any kind of fibrous ‘pipe jointing’ shouldbe absolutely forbidden as fibre contamination is difficult to clear.

Preferably, all fuel lines, and certainly all underground lines, should be constructedwith orbital welded joints, or with the use of compression fittings approved forthe fuels concerned (some fuels such as ‘winter diesel’ appear to be particularlypenetrating). External fuel oil lines should be lagged and must be trace heated iftemperatures are likely to fall to a level at which fuel ‘waxing’ may take place.

Within the test cell, flexible lines may be required in short sections to allowfor engine rigging or connection to fixed instrumentation; such tubing and fittingsmust be specifically made and certified to handle the range of fuels being used;the generally recommended specification is for metal braided, electrically conductive,Teflon hoses.

Storage and treatment of residual fuels

Many large stationary diesel engines and the majority of slow speed marine enginesoperate on heavy residual fuels that require special treatment before use and mustalso be heated before delivery to the fuel injection system. In such cases, the test cellfuel supply system is required to incorporate the special features of the fuel supplyand treatment systems installed in diesel propelled ships. The bulk storage tanks willrequire some form of heating coils within it to enable the heavy oil to be pumpedinto a heated settling tank.

It is always necessary with fuels of this type to remove sludge and water beforeuse in the engine. The problem here is that the density of residual fuels can approachthat of water, making separation very difficult. The accepted procedure is to raise thetemperature of the oil, thus reducing its density, and then to feed it through a purifierand clarifier in series. These are centrifugal devices, the first of which removes mostof the water, while the second completes the cleaning process.

Refs 1 and 2 give detailed recommendations for the design and operation of suchsystems and Fig. 7.3 shows a schematic arrangement. It is also necessary to providechangeover arrangements in the test cell so that the engine can be started and shutdown on a light fuel oil.


Settlingtank

Preheater

Feed pump

Q

Purifier Clarifier

Sludge

Water

Daytank

Sludge

Figure 7.3 Fuel supply and treatment system for residual fuel oil

Storage of biofuels

Test facilities involved in the preparation and mixing of biofuels may have specialstorage problems to consider. Most biodiesel fuels are a mixture of petroleum andvegetable-derived liquids; they are usually designated by the prefix B followed bythe percentage of the biomass derived content; thus B20 contains 20 per cent ofnon-petroleum fuel. Biofuels tend to be hydrophilic, therefore storage tanks and linescan be subject to corrosion from condensation.

Laboratories working with the constituents of biodiesel mixtures will requirestorage for animal- or vegetable-derived oils in bulk or drums; these may requirelow-grade heating and stirring in some climatic conditions. Water in the fuel willtend to encourage the growth of microbe colonies in heated fuel tanks which canform soft masses that plug filters.

Ethanol/gasoline mixtures are designated with the prefix E followed by thepercentage of ethanol. E5 and E10 (commonly called ‘gasohol’) are commerciallyavailable in various parts of Europe. E20 and E25 are standard fuel mixes in Brazilfrom where no particular technical problems of storage have been reported.

The storage and handling of 100 per cent ethanol and methanol raises problemsof security; the former is highly intoxicating and the latter highly toxic and anyonelikely to drink the intoxicant probably lacks the chemical knowledge to distinguishbetween the two.

Underground fuel lines

Buried mild steel fuel lines, particularly those run under and through concrete floorslabs, should be wrapped with water-repellent bandage (in the UK, ‘Denzo tape’)and laid on fine gravel within a well-compacted trench. It is good practice to rununderground fuel lines in a sealed trench of concrete sections with a load-bearinglid; such ducts should be fitted with hydrocarbon ‘sniffers’ connected to an alarm

134 Engine Testing

system so as to detect fuel leakage before ground contamination takes place. Thereare certain special requirements for the fuel supply system for an engine test facilitythat may not apply to systems for other purposes; these include the provision forthe use of a number of different fuels including the same quantities of referencefuels. The ‘fuel farm’ for a modern R&D facility may be provided with a number ofstorage tanks and several ring mains; these must be clearly labelled at all filling anddraw-off points.

Reference fuel drums

Where frequent use is made of special or reference fuels, supplied by the drum, specialprovision must be made for their transportation from the secure fuel store, such asthat shown in Fig. 7.2, and protection while in use. Special drum containers, designedto be transported by fork-lift, are available and these may be parked immediatelyagainst an outside wall of the test cell requiring the supply. Connection to the cellsystem can be achieved by using an automotive 12V pump system designed to screwinto the drum and deliver fuel through a connection point outside the cell fitted withidentical fire isolation interlocks as used in the permanently plumbed supplies.

Natural gas, liquefied natural gas, compressed natural gas(NG, LNG, CNG)

Natural gas consists of about 90 per cent methane, which has a boiling point atatmospheric pressure of −163�C. Engine test installations requiring natural gas fordual-fuel engines usually draw this from a mains supply at just above atmosphericpressure so high-pressure storage arrangements are not necessary. In the UK, thedistribution of LNG to individual commercial users is via a national grid and generallycovered by the Gas (Third Party Access) Regulations 2004. Fire hazards are moderatewhen compared with LPG installations.

Engine test cells using LNG have to comply with gas industry regulations whichmay include explosion relief panels in the cell and building exhaust system.

Fuel supply to the test cell

Fuel supplies to a cell may be provided either under static head from a day tank,a pumped supply from a reference fuel drum (see above) or by a pressurized reticu-lation system fed from the central fuel farm (Fig. 7.4).

Day tanks should be fuel specific to prevent cross-contamination, so there maybe a requirement for several. Modern safety practice usually, but not exclusively,dictates that gasoline day tanks should be kept on the outside wall of the test facility.


Fuel tank

Fuel supplypump

Cellsupply

Fire N/Csolenoid valve

No horizontalpipes

Fire/safetydumpvalve

Static head≈11.8 m per/barfor fuel oil

Overflow to sump

Vent to atmosphereFloat or levelpump-switch

High alarm

Low alarm

Figure 7.4 A schematic showing the elements of a typical fuel day tank system

In all cases, day tanks must be fitted with a dump valve for operation in caseof fire; this allows fuel contained in the vulnerable above-ground tanks and pipesto be returned to the fuel farm, or a specific underground dump tank. In addition tohaving to be vented to atmosphere, day tanks also have to be fitted with a monitoredoverflow and spill return system in the case of a malfunction of any part of thesupply system.

136 Engine Testing

The static head is commonly at 4.5m or above, but may need to be calculatedspecifically in order to achieve the 0.5–0.8 bar inlet pressure required by someindustrial standard fuel consumption and treatment instruments.

When the tanks are exposed to ambient weather conditions, they should be shieldedfrom direct sunlight and, in the case of diesel fuels, lagged or trace-heated.

If a pumped system is used, it must be remembered that for much of the operatinglife, the fuel demand from the cells will be below the full rated flow of the pump,therefore the system must be able to operate under bypass, or stall, conditions withoutcavitation or undue heating.

The use of positive displacement, pneumatically operated pumps, incorporatinga rubberized air/fluid diaphragm, have a number of practical advantages: they donot require an electrical supply and they are designed to be able to stall at full(regulated air) pressure, thus maintaining a constant fuel pressure supply to the testcells.

In designing systems to meet this wide range of requirements, some generalprinciples should be borne in mind:

• Pumps handling gasoline should have a positive static suction head to preventcavitation problems on the suction side.

• Fuel lines (pipes) with low flow resistance where bends rather than sharp elbowsare used and no sudden changes in internal cross-section exist.

• Fuel lines should be pressure tested by competent staff prior to filling with fuel.• Each fuel line penetrating the cell wall should be provided with a normally closed

solenoid operated valve interlocked with both the cell control system and the fireprotection circuits

In-cell fuel systems

The fuel system in the test cell will vary widely in complexity. In some cases thesystem may be limited to a single fuel line connected to the engine’s fuel pump, buta special purpose test cell may call for the supply of many different fuels all passingthrough fuel temperature control and fuel consumption measurement devices.

For the occasional test, a simple fuel tank of the type used for outboard engines,capacity not more than 10 litres, may be all that is necessary. These marine devicesare certified as safe to use in their designated roles of containing and supplying fuelto an engine; the use of other containers is not safe and would endanger the insuranceof the premises in which they are used.

The following points should be considered in the planning of in-cell systems:

• It may be necessary to provide several separate fuel supplies to a cell; a typicalprovision would be three lines for diesel fuel, ‘standard’ and ‘super’ unleadedgasoline, respectively. Problems can arise from carry-over, with consequentdanger of ‘poisoning’ exhaust catalysers. The capacity of the fuel held in thecell system, including such items as filters, needs to be kept to a minimum.


To minimize cross-contamination it is desirable although often difficult to locatethe common connection as close to the engine as possible. A common layoutprovides for an inlet manifold of fuels to be fitted below the in-cell fuelconditioning and measuring systems. Selection of fuel can be achieved via aflexible line fitted with a self-sealing connector from the common system to thedesired manifold mounted connector.

• It is good practice to have a cumulative fuel meter in each line for general auditand for contract charging.

• Air entrainment and vapour locking can be a problem in test cell fuel systems.An air eliminating valve should be fitted at the highest point in the system, withan unrestricted vent to atmosphere external to the cell and at a height and in aposition that prevents fuel escape.

• Breaking of fuel lines, with consequent spillage, should as far as possiblebe avoided. It is sensible to mount all control components on a permanentwall-mounted panel or within a special casing, with switching via interlockedand clearly marked valves. The run of the final flexible fuel lines to the engineshould not interfere with operator access. A common arrangement is to run thelines via an overhead boom. Self-sealing couplings should be used for engineand other frequently broken connections.

• It is essential to fit oil traps to cell drain connections to avoid the possibility ofdischarging oil or fuel into the foul water drains.

• Aim for a flow rate during normal operation to be between 0.2 and a maximumrecommended fuel line velocity: 1.0m/s.

Engine fuel pressure control

There are three fuel pressure control problems that may require to be solved whendesigning or operating an engine test cell:

1. Pressure of fuel supply to the fuel conditioning system or consumptionmeasurement instrument. Devices such as the AVL Fuel Balance require amaximum pressure of 0.8 bar at the instrument inlet; in the case of pumpedsystems this may require a pressure reducing regulator to be fitted before theinstrument.

2. Pressure of fuel supplied to the engine. Typically systems fitted in normalautomotive cells are adjustable to give pressures at the engine system inlet ofbetween 0.05 and 4 bar.

3. Pressure of the fuel being returned from the engine. When connected to a fuelconditioning and consumption system, the fuel return line in the test cell maycreate a greater back pressure than required. In the vehicle, the return linepressure may be between zero and 0.5 bar, whereas cell systems may requireover 1.5 bar to force fuel through conditioning equipment and to overcomestatic head differences between the engine and instruments, therefore a pressurereducing circuit including a pump may be required. Figure 7.5 shows a circuit

138 Engine Testing

From engine Out

Engine return

Engine feed In

Out

2

1

Figure 7.5 Regulation of engine supply and return pressures within dotted lines(1) is the fuel return pressure control and (2) the fuel feed to the engine fromthe fuel consumption measurement system (AVL List)

of a system that allows for the independent regulation of engine supply andreturn pressures. Care must be exercised in the design of such circuits so as toavoid the fuel pressure being taken below the point at which vapour bubblesform.

Engine fuel temperature control

Clearly all the materials used in the system used in control of fuel temperature,such as heat exchangers, piping and sealing materials, must be checked with themanufacturer as to their suitability for use with all specified fuels.

If fuel supplied to the engine has to be maintained at one standard temperature,as is usually the case in quality audit cells, then a relatively simple control systemmay provide hot water circulated at controlled temperature through the water-to-fuelheat exchanger. Such an arrangement can give good temperature control despite widevariations in fuel flow rate.

The control of the fluid fuel temperature within the engine circuit is com-plicated by the fuel rail and spill-back strategy adopted by the engine designer,because of this and variations in the engine rigging pipe work, no one circuit designcan be recommended. Commercially available fuel conditioning devices can onlyspecify the temperature control at the unit discharge: the system integrator has toensure the heat gains and losses within the connecting pipe work do not invalidateexperiments.

It is particularly important to minimize the distance between the temperaturecontrolling element and the engine so that, if running is interrupted, the enginereceives fuel at the desired temperature with the minimum of delay.


Commercially available units typically have the following basic specifications:

• setting fuel temperature: upto 80�C; providing application system is designed andpressurized to prevent excessive vaporization;

• deviation between set value and actual value at instrument output: <1�C;• power input: 0.4 kW or 2 kW (with heating);• cooling load: typically 1.5 kW.

It should be noted that the stability of temperature control below ambient will bedependent on the stability of the chilled water supply which should have a controlloop independent of other, larger, cooling loads in any shared system.

A constant fuel temperature over the duration of an experiment is a prerequisitefor accurate consumption measurement (see Chapter 12).

As with all systems covered in this book, the complexity and cost of fueltemperature control will depend on the operation range and accuracy specified.

To confirm that temperature change across the fuel measurement circuit hasa significant influence on fuel consumption and fuel consumption measurement,especially during low engine power periods Fig. 7.6 should be considered. Thegraph lines of Fig. 7.6 show that by making a cyclic step change in the set pointof a fuel temperature controller of only ±0.2�C within the measurement circuit,

2.75

2.7

2.65

2.6

2.55

2.5

2.45200 210 220 230 240 250

Time, s

Actual temperature T2

Set temperature

Temperature at engine inlet T3

Fuel consumption

Fuel consum

ption

, kg/h

260 270 280 290

~25 s

300

+/– 0.1°C

+/– 0.2°C

+/– 0.05 kg/h

Tem

pera

ture

,°C

15.5

16

16.5

17

17.5

18

18.5

Figure 7.6 Effects of fuel temperature changes on the fuel consumptionmeasurement values. T2 is the controlled outlet of the fuel conditioning devicesuperimposed on the step demand changes. T3 is the temperature at theengine inlet. The difference between the two is due to system damping(AVL List)

140 Engine Testing

which produces an oscillation in fuel temperature at the engine inlet of ±0.1�C,then a variation in the measured fuel consumption value of ±0.05 kg/h (top line) iscreated.

This equals a variation of ±2 per cent of the actual fuel consumption value andshows that instrument manufacturers’ claims of fuel consumption accuracy are onlyvalid under conditions of absolutely stable fuel temperatures.

Engine oil cooling systems

Certain special requirements apply to oil cooling units. Unless the cell is operatinga ‘dry-sump’ lubrication system on the engine, the entire lubrication oil temperaturecontrol circuit must lie below engine sump level so that there is no risk of floodingthe sump. It may be necessary to provide heaters in the circuit for rapid warm-up ofthe engine; in this case the device must be designed in such a way as to ensure thatthe skin temperature of the heating element cannot reach temperatures at which oil‘cracking’ an occur.

Figure 7.7 shows schematically a separate lubricating oil cooling and conditioningunit.

Where very accurate transient temperature control is necessary, the use of separatepallet-mounted cooling modules located close to the engine may be required;otherwise permanently located oil and cooling water modules offer the best solutionfor most engine testing. One common position is behind the dynamometer whereboth units may be fed from the external cooling water system and the engine con-nection hoses may be run under the dynamometer to a connection point near theshaft end.

To engine

By pass

Fromengine

Pressureregulator Heater

Pump

Cooler

Coolingwater in

Control valve

Figure 7.7 Schematic of oil temperature control unit with sensing points andcontrol connections omitted


Properties of gasoline and ‘shelf life’

Automotive fuels as sold at the retail pump in any country of the world are variablein many details, some of which are appropriate to the geographic location and seasonand others are due to the original feedstock from which the fuel was refined. Addedto the experimental variable of retail fuel properties the storage of gasoline in apartially filled container in conditions where it is exposed to direct sunlight will causeit to degrade rapidly, further increasing the variables in any engine testing. Referencefuels of all types should be stored in dark cool conditions and in containers that are90–95 per cent full; their shelf life should not be assumed to be more than a fewmonths.

A very important fuel property is the calorific value. Engine development workis often concerned with ‘chasing’ very small improvements in fuel consumption,differences that can easily be swamped by variations in the calorific value of thefuels. Similarly, comparisons of identical engines manufactured at different sites willbe invalid if they are tested on different fuels or fuels that have been allowed todeteriorate at a differential rates.

For example, the LCVs of hexane and benzene, typical constituents of gasoline,are, respectively, 44.8 and 40.2MJ/kg. A typical value for gasoline would be43.9MJ/kg, but this could easily vary by ±2 per cent in different parts of theworld, while the presence of alcohol as a constituent can depress the calorific valuesubstantially.

Octane number is the single most important gasoline specification, since it governsthe onset of detonation or ‘knock’ in the engine. This condition, if allowed tocontinue, will rapidly destroy an engine. Knock limits the power, compression ratioand hence the fuel economy of an engine. Too low an octane number also gives riseto run-on when the engine is switched off.

Three versions of the octane number are used: research octane number (RON),motor octane number (MON) and front end octane number (R100). RON isdetermined in a specially designed (American) research engine: the Co-operativeFuel Research (CFR) engine. In this engine, the knock susceptibility of the fuel isgraded by matching the performance with a mixture of reference fuels, iso-octanewith a RON= 100 and n-heptane with a RON= 0.

The RON test conditions are now rather mild where modern engines are concernedand the MON test imposes more severe conditions but also uses the CFR engine.The difference between the RON and the MON for a given fuel is a measure ofits ‘sensitivity’. This can range from about 2 to 12, depending on the nature of thecrude and the distillation process. In the UK, it is usual to specify RON, while in theUSA the average of the RON and the (lower) MON is preferred. R100 is the RONdetermined for the fuel fraction boiling at below 100�C.

Volatility is the next most important property and is a compromise. Low volatilityleads to low evaporative losses, better hot start, less vapour lock and less carburettoricing. High volatility leads to better cold starting, faster warm-up and hence bettershort-trip economy, also to smoother acceleration.

142 Engine Testing

Mixture preparation

Ideally, each cylinder of a multicylinder engine should receive under all operatingconditions exactly the same charge of fuel and air, in exactly the same condition,as every other cylinder. This is an unattainable ideal, but a large element in thedevelopment programme of any new engine is concerned with approaching it asclosely as possible.

A principal variable is the method of introducing the fuel. In order of increasingcost, there are essentially three solutions: carburettor, single (throttle body) injectorand individual port injectors. The main factors to be considered are:

• Equal distribution of air quantity, involving detail design of manifold andinteraction with flow leaving throttle or carburettor. Particularly difficult at partthrottle.

• Equal distribution of fuel. With carburettor or single-point injection there are threeaspects to be considered: distribution of air, of fuel droplets and of fuel vapour.The proportion of the latter increases as power is reduced with accompanyingincreased manifold depression.

• Deposition of liquid fuel on manifold walls.• Nature of mixture: degree of atomization and of vapour formation.

Properties of diesel fuels

Just as the range of size of the diesel engine is much greater than that of the spark igni-tion engine, from 1–2 kW to 50 000 kW, the range of fuel quality is correspondinglygreat (see Chapter 17 for main relevant specifications).

Cetane number is the most important diesel fuel specification. It is an indicationof the extent of ignition delay: the higher the cetane number, the shorter the ignitiondelay, the smoother the combustion and the cleaner the exhaust. Cold starting is alsoeasier the higher the cetane number.

Viscosity covers an extremely wide range. BS 2869 specifies two grades of vehicleengine fuels, class Al and class A2, having viscosities in the range 1.5–5.0 cSt and1.5–5.5 cSt, respectively, at 40�C. BS MA 100 deals with fuels for marine engines.It specifies nine grades, class M1, equivalent to class A1, with increasing viscositiesup to class M9, which has a viscosity of 130 cSt max at 80�C.

Summary

The bulk storage and handling of automotive fuels has been briefly described andattention drawn to the high degree of safety and environmental legislation that isinvolved in the design of such systems. A specialist contractor is recommended


for any work involving volatile liquid fuels and is required by regulation in manycountries when installing gaseous fuel systems.

Liquid fuels are complex mixtures and can be significantly variable; therefore,the need for reference fuels has been discussed as has the possible deterioration offuels when stored in poor conditions.

The requirement for accurate fuel pressure and temperature control in theengine cell has been demonstrated. For coverage of fuel consumption methods, seeChapter 12; a list of useful publications is to be found below.

References

1. Hughes, J.R. and Swindells, S. (1987) The Storage and Handling of Petroleum Liquids,Griffin, London.

2. Recommendations for Cleaning and Pretreatment of Heavy Fuel Oil, Alfa Laval, London.

Further reading

BS 3016 Specification for Pressure Regulators and Automatic Changeover Devices

for Liquefied Petroleum Gases.BS 4040 Specification for Leaded Petrol (Gasoline) for Motor Vehicles.BS 4250 Part 1 Specification for Commercial Butane and Propane.BS EN589 Specification for Automotive LPG.BS 5355 Specification for Filling Ratios and Developed Pressures for Liquefiable

and Permanent Gases.BS 6843 Parts 0 to 3 Classification of Petroleum Fuels.BS EN228 Specification of Unleaded Petrol (Gasolene) for Motor Vehicles.BS 7405 Guide to Selection and Application of Flowmeters for the Measurement of

Fluid Flow in Closed Conduits.Dangerous Substances and Explosive Atmospheres Regulations (DSEAR) 2002.

BS 2869 Part 2 Specification for Fuel Oil for Agricultural and Industrial Engines

and Burners.Installation, Decommissioning and Removal of Underground Storage Tanks: PPG27,

UK Environmental Agency.UK Guidance and rules for the storage of gasoline in portable containers can be

found at http://www.hse.gov.uk/lau/lacs/65–9.htm

8 Dynamometers and themeasurement of torque

Introduction

The torque produced by a prime mover under test is resisted and measured by thedynamometer to which it is connected. The accuracy with which a dynamometermeasures both torque and speed is fundamental to all the other derived measurementsmade in the test cell.

In this chapter the principles of torque measurement are reviewed and then thetypes of dynamometer are reviewed in order to assist the purchaser in the selectionof the most appropriate machine.

Measurement of torque: trunnion-mounted (cradle) machines

The essential feature of trunnion-mounted or cradled dynamometers is that the powerabsorbing element of the machine is mounted on bearings coaxial with the machineshaft and the torque is restrained and measured by some kind of transducer actingtangentially at a known radius from the machine axis.

Until the beginning of the present century, the great majority of new and existingdynamometers used this method of torque measurement. In traditional machines thetorque measurement was achieved by physically balancing a combination of deadweights and a spring balance against the torque absorbed (Fig. 8.1). As the stiffnessof the balance was limited, it was necessary to adjust its position depending on thetorque, to ensure that the force measured was accurately tangential.

Modern trunnion-mounted machines, shown diagrammatically in Fig. 8.2, usea force transducer, almost invariably of the strain gauge type, together with anappropriate bridge circuit and amplifier. The strain gauge transducer or ‘load cell’ hasthe advantage of being extremely stiff, so that no positional adjustment is necessary,but the disadvantage of a finite fatigue life after a (very large) number of loadapplications. The backlash and ‘stiction’-free mounting of the transducer betweencarcase and base is absolutely critical.

The trunnion bearings are either a combination of a ball bearing (for axial location)and a roller bearing or hydrostatic type. These bearings operate under unfavourableconditions, with no perceptible angular movement, and the rolling element type is

Dynamometers and the measurement of torque 145

T

RF

W

Figure 8.1 Diagram of Froude type, trunnion-mounted, sluice-gatedynamometer measuring torque with dead weights and spring balance

consequently prone to brinelling, or local indentation of the races, and to fretting. Thisis aggravated by vibration that may be transmitted from the engine and periodicalinspection and turning of the outer bearing race is recommended in order to avoidpoor calibration. A Schenck dynamometer design (Fig. 8.3) replaces the trunnionbearings by two radial flexures, thus eliminating possible friction and wear, but at theexpense of the introduction of torsional stiffness, of reduced capacity to withstandaxial loads and of possible ambiguity regarding the true centre of rotation, particularlyunder side loading.

Measurement of torque using in-line shafts or torque flanges

A torque shaft dynamometer is mounted in the drive shaft between engine and brakedevice. It consists essentially of a flanged torque shaft fitted with strain gauges anddesigns are available both with slip rings and with RF signal transmission. Figure 8.4is a brushless torque shaft unit intended for rigid mounting.

More common in automotive testing is the ‘disc’ type torque transducer, com-monly known as a torque flange (Fig. 8.5), which is a device that is bolted directlyto the input flange of the brake and transmits data to a static antenna encircling it.

146 Engine Testing

R

T

F

Carcase

Fixed

base

Figure 8.2 Diagram of trunnion-mounted dynamometer measuring torque witha load cell

A perceived advantage of the in-line torque measurement arrangement is that itavoids the necessity, discussed below, of applying torque corrections under transientconditions of torque measurement. However, not only are such corrections, usingknow constants, trivial with modern computer control systems, there are importantproblems that may reduce the inherent accuracy of this arrangement.

For steady state testing, a well-designed and maintained trunnion machine willgive more consistently auditable and accurate torque measurements than the inlinesystems; the justification for this statement can be listed as follows:

• The in-line torque sensor has to be oversized for the rating of its dynamometerand being oversized the resolution of the signal is lower. The transducer hasto be overrated because it has to be capable of dealing with the instantaneoustorque peaks of the engine which are not experienced by the load cell of atrunnion-bearing machine.

• The transducer forms part of the drive line and requires very careful installationto avoid the imposition of bending or axial stresses on the torsion sensing elementfrom other components or its own clamping device.

• The in-line device is difficult to protect from temperature fluctuations within andaround the drive-line.

• Calibration checking of these devices is not as easy as for a trunnion-mountedmachine; it requires a means of locking the dynamometer shaft in addition tothe fixing of a calibration arm in a horizontal position without imposing bendingstresses.

• Unlike the cradled machine and load cell, it is not possible to verify the measuredtorque of an in-line device during operation.


10

9 4 8 1 5 7 2 3

11 12 6 13 14

Figure 8.3 Schenck, dry gap, disc type eddy-current dynamometer 1, rotor; 2,rotor shaft; 3, coupling flange; 4, water outlet with thermostat; 5, excitation coil;6, dynamometer housing; 7, cooling chamber; 8, air gap; 9, speed pick-up; 10,flexure support; 11, base; 12, water inlet; 13, joint; 14, water outlet pipe

148 Engine Testing

A = Mounting flange

A 3 4 5 6 7 B

1 1 = Torsion element (rotor)2 = Applied SGs3 = Spindle bearing4 = Housing (stator)5 = Elastic seal6 = Capacitive transmission7 = Inductive transmission8 = Toothed ring for speed measurement9 = Speed pick-up

10 = Cable connection box

2

8

9

10

B = Flange for torque introduction

Figure 8.4 Brushless torque-shaft for mounting in shaft-line between engineand ‘brake’

Antenna segments

Rotor

Adaptor flangeMeasuring body

Figure 8.5 Shaft-line components of a torque flange

It should be noted that, in the case of modern a.c. dynamometer systems, the tasksof torque measurement and torque control may use different data acquisition paths.In some installations the control of the trunnion-mounted machine may use its owntorque calculation and control system, while the test values are taken from an inlinetransducer such as a torque shaft.


Calibration and the assessment of errors in torquemeasurement

We have seen that in a conventional dynamometer, torque T is measured as a productof torque arm radius R and transducer force F.

Calibration is invariably performed by means of a calibration arm, supplied by themanufacturer, which is bolted to the dynamometer carcase and carries dead weightswhich apply a load at a certified radius. The manufacturer certifies the distancebetween the axis of the weight hanger bearing and an axis defined by a line joiningthe centres of the trunnion bearings (not the axis of the dynamometer, which indeedneed not precisely coincide with the axis of the trunnions).

There is no way, apart from building an elaborate fixture, in which the dynamome-ter user can check the accuracy of this dimension: he is entirely in the hands ofthe manufacturer. The arm should be stamped with its effective length. For R&Dmachines of high accuracy the arm should be stamped for the specific machine.

The ‘dead weights’ should in fact be more correctly termed ‘standard masses’.They should be certified by an appropriate standards authority located as near aspossible to the geographical location in which they are used. The force they exerton the calibration arm is the product of their mass and the local value of ‘g’. Thisis usually assumed to be 9.81m/s2 and constant: in fact this value is only correctat sea level and a latitude of about 47� N. It increases towards the poles and fallstowards the equator, with local variations. As an example, a machine calibrated inLondon, where g = 9.81m/s2 , will read 0.13 per cent high if recalibrated in Sydney,Australia and 0.09 per cent low if recalibrated in St Petersburg without correctingfor the different local values of g.

These are not negligible variations if one is hoping for accuracies better than1 per cent. The actual process of calibrating a dynamometer with dead weights, iftreated rigorously, is not entirely straightforward. We are confronted with the factsthat no transducer is perfectly linear in its response, and no linkage is perfectlyfrictionless. We are then faced with the problem of adjusting the system so as toensure that the (inevitable) errors are at a minimum throughout the range.

A suitable calibration procedure for a machine using a typical strain-gauge loadcell for torque measurement is as follows.

The dynamometer should not be coupled to the engine. After the system has beenenergized long enough to warm up the load cell output is zeroed with the machinein its normal no-load running condition (cooling water on, etc.) and the calibrationarm weight balanced by equal and opposite force. Dead weights are then added toproduce approximately the rated maximum torque of the machine. This torque iscalculated and the digital indicator set to this value.

The weights are removed, the zero reading noted, and weights are added, prefer-ably in 10 equal increments, the cell readings being noted. The weights are removedin reverse order and the readings again noted.

The procedure described above means that the load cell indicator was set to readzero before any load was applied (it did not necessarily read zero after the weights

150 Engine Testing

had been added and removed), while it was adjusted to read the correct maximumtorque when the appropriate weights had been added.

We now ask: is this setting of the load cell indicator the one that will minimizeerrors throughout the range and are the results within the limits of accuracy claimedby the manufacturer?

Let us assume we apply this procedure to a machine having a nominal rating of600Nm torque and that we have six equal weights, each calculated to impose a torqueof 100Nm on the calibration arm. Table 8.1 shows the indicated torque readings forboth increasing and decreasing loads, together with the calculated torques appliedby the weights. The corresponding errors, or the differences between torque appliedby the calibration weights and the indicated torque readings are plotted in Figs 8.6and 8.7.

The machine is claimed to be accurate to within ±0.25 per cent of nominal ratingand these limits are shown. It will be clear that the machine meets the claimed limitsof accuracy and may be regarded as satisfactorily calibrated.

It is usually assumed, though it is not necessarily the case, that hysteresis effects,manifested as differences between observed torque with rising load and with fallingload, are eliminated when the machine is running, due to vibration, and it is acommon practice when calibrating to knock the machine carcase lightly with a softmallet after each load change to achieve the same result.

It is certainly not wise to assume that the ball joints invariably used in the calibra-tion arm and torque transducer links are frictionless. These bearings are designed forworking pressures on the projected area of the contact in the range 15 to 20MN/m2

and a ‘stick slip’ coefficient of friction at the ball surface of, at a minimum, 0.1 isto be expected. This clearly affects the effective arm length (in either direction) andmust be relaxed by vibration.

Some large dynamometers are fitted with torque multiplication levers, reducingthe size of the calibration masses. In increasingly litigious times and ever morestringent health and safety legislation, the frequent handling of multiple 20 or 25 kg

Table 8.1 Dynamometer calibration (example taken from actual machine)

Mass

(kg)

Applied torque

(Nm)

Reading

(Nm)

Error (Nm) Error

(% reading)

Error

(% full scale)

0 0 0�0 0�0 0�0 0�010 100 99�5 −0�5 −0�5 −0�08330 300 299�0 −1�0 −0�33 −0�16750 500 500�0 0�0 0�0 0�060 600 600�0 0�0 0�0 0�040 400 400�5 +0�5 +0�125 +0�08320 200 200�0 0�0 0�0 0�00 0 0�0 0�0 0�0 0�0


1.000

0.800

0.600

0.400

0.200

0.000

–0.200

–0.400

–0.600

–0.800

–1.000

Applied torque Nm

0.000 100.000 200.000 300.000 400.000 500.000 600.000 700.000

% E

rro

r

Figure 8.6 Dynamometer calibration error as percentage of applied torque

0.250

0.200

0.150

0.100

0.050

0.000

–0.050

–0.100

–0.150

–0.200

–1.250

Applied torque

% E

rro

r

0.000 100.000 200.000 300.000 400.000 500.000 600.000 700.000

Figure 8.7 Dynamometer calibration error as percentage of full scale

weights may not be advisable. It is possible to carry out torque calibration byway of ‘master’ load cells or proving rings.∗ These devices have to be mounted ina jig attached to the dynamometer and give an auditable measurement of the forcebeing applied on the target load cell by means of a hydraulic actuator. Such systemsproduce a more complex ‘audit trail’ in order to refer the calibration back to nationalstandards.

It is important when calibrating an eddy-current machine that the water pressurein the casing should be at operational level, since pressure in the transfer pipes cangive rise to a parasitic torque. Similarly, any disturbance to the run of electrical

∗ A proving ring is a hollow steel alloy ring whose distortion under a rated range of com-pressive loads is known and measured by means of an internal gauge.

152 Engine Testing

cables to the machine must be avoided once calibration is completed. Finally, it ispossible, particularly with electrical dynamometers with forced cooling, to developsmall parasitic torques due to air discharged non-radially from the casing. It is aneasy matter to check this by running the machine uncoupled under its own powerand noting any change in indicated torque.

Experience shows that a high grade dynamometer such as would be used forresearch work, after careful calibration, may be expected to give a torque indicationthat does not differ from the absolute value by more than about ±0.1 per cent of thefull load torque rating of the machine.

Systematic errors such as inaccuracy of torque arm length or wrong assumptionsregarding the value of g will certainly diminish as the torque is reduced, but othererrors will be little affected: it is safer to assume a band of uncertainty of constantwidth. This implies, for example, that a machine rated at 400Nm torque with anaccuracy of ±0.25 per cent will have an error band of ±1N. At 10 per cent of ratedtorque, this implies that the true value may lie between 39 and 41Nm. It is as well tomatch the size of the dynamometer as closely as possible with the rating of the engine.

All load cells used by reputable dynamometer manufacturers will compensate forchanges in temperature, though their rate of response to a change may vary. Theywill not, however, be able to compensate for internal temperature gradients induced,for example, by air blasts from ventilation fans or radiant heat from exhaust pipes.

The subject of calibration and accuracy of dynamometer torque measurement hasbeen dealt with in some detail, but this is probably the most critical measurementthat the test engineer is called upon to make, and one for which a high standardof accuracy is expected but not easily achieved. Calibration and certification of thedynamometer and its associated system should be carried out at the very least oncea year, and following any system change or major component replacement.

Torque measurement under accelerating and deceleratingconditions

With the increasing interest in transient testing it is essential to be aware of the effectof speed changes on the ‘apparent’ torque measured by a trunnion-mounted machine.

The basic principle is simple:

Inertia of dynamometer rotor I kg m2

Rate of increase in speed � rad/s2

N rpm/sInput torque to dynamometer T1 NmTorque registered by dynamometer T2 Nm

T1−T2 = I�=2�NI

60Nm

= 0�1047NI Nm


To illustrate the significance of this correction, a typical eddy-current dynamome-ter capable of absorbing 150 kW with a maximum torque of 500Nm has a rotorinertia of 0.11 kg m2. A d.c. regenerative machine of equivalent rating has a rotationalinertia of 0.60 kg m2.

If these machines are coupled to an engine that is accelerating at the comparativelyslow rate of 100 rpm/s the first machine will read the torque low during the transient

phase by an amount:

T1−T2 = 0�1047×100×0�11= 1�15 Nm

while the second will read low by 6.3Nm.If the engine is decelerating, the machines will read high by the equivalent amount.Much larger rates of speed change are demanded in some transient test sequences

and this can represent a serious variation of torque indication, particularly whenusing high inertia dynamometers.

With modern computer processing of the data, corrections for these and otherelectrically induced transient effects can be made with software supplied by test plantmanufacturers.

Measurement of rotational speed

Rotational speed of the dynamometer is measured either by a system using a toothedwheel and a pulse sensor within its associated electronics and display or, morerecently, by use of an optical encoder system. While the pulse pick-up system isrobust and, providing the wheel to transducer gap is correctly set and maintained,reliable, the optical encoders, which use the sensing of very fine lines etched on asmall disk, need more care in mounting and operation. Since the commonly usedoptical encoders transmit over 1000 pulses per revolution, misalignment of its drivemay show up as a sinusoidal speed change, therefore they are normally mounted aspart of an accurately machined assembly forming part of the machine housing.

It should be remembered that with bidirectional dynamometers and modern electri-cal machines operating in four quadrants (Fig. 8.8), it is necessary to measure not onlyspeed but also direction of rotation. Encoder systems can use separate tracks of theirengraved disks to sense rotational direction. It is extremely important that the opera-tor uses a common and clearly understood convention describing direction of rotationthroughout the facility, particularly in laboratories operating reversible prime movers.

As with torque measurement, specialized instrumentation systems may use sepa-rate transducers for the measurement of speed or for the control of the dynamometer.In many cases, engine speed is monitored separately and in addition to dynamometerspeed. The control system can use these two signals to shut down automatically inthe case of a shaft failure. Measurement of crank position during rotation is discussedin Chapter 14.

154 Engine Testing

Torque

2

Clockwise,

absorb

torque

Clockwise,

develop

torque

Rotation

Anticlockwise,

absorb

torque

Anticlockwise,

develop

torque

1

3 4

Figure 8.8 Dynamometer operating quadrants

Measurement of power, which is the product of torque and speed, raises theimportant question of sampling time. Engines never run totally steadily and thetorque transducer and speed signals invariably fluctuate. An instantaneous reading ofspeed will not necessarily, or even probably, be identical with a longer-term average.Choice of sampling time and of the number of samples to be averaged is a matter ofexperimental design and compromise.

Choice of dynamometer

Perhaps the most difficult question facing the engineer setting up a test facilityis the choice of the most suitable dynamometer. In this part of the chapter thecharacteristics, advantages and disadvantages of the various types are discussed anda procedure for arriving at the correct choice is described.

The earliest form of dynamometer, the rope brake dates back to the early yearsof the last century. An extremely dangerous device, it was nevertheless capable ofgiving quite accurate measurements of power. Its successor, the Prony brake, alsorelied on mechanical friction and like the rope brake required cooling by waterintroduced into the hollow brake drum and removed by a scoop.

Both these devices are only of historical interest. Their successors may beclassified according to the means adopted for absorbing the mechanical power of theprime mover driving the dynamometer.

Classification of dynamometers

1. Hydrokinetic or ‘hydraulic’ dynamometers (water brakes). With the exceptionof the disc dynamometer, all machines work on similar principles (Fig. 8.9). A shaftcarries a cylindrical rotor which revolves in a watertight casing. Toroidal recesses


formed half in the rotor and half in the casing or stator are divided into pocketsby radial vanes set at an angle to the axis of the rotor. When the rotor is driven,centrifugal force sets up an intensive toroidal circulation as indicated by the arrowsin Fig. 8.9a. The effect is to transfer momentum from rotor to stator and hence todevelop a torque resistant to the rotation of the shaft, balanced by an equal andopposite torque reaction on the casing.

A forced vortex of toroidal form is generated as a consequence of this motion,leading to high rates of turbulent shear in the water and the dissipation of power inthe form of heat to the water. The centre of the vortex is vented to atmosphere byway of passages in the rotor and the virtue of the design is that power is absorbedwith minimal damage to the moving surfaces, either from erosion or from the effectsof cavitation.

The machines are of two kinds, depending on the means by which the resistingtorque is varied.

1(a) Constant fill machines: the classical Froude or sluice plate design, Fig. 8.10.In this machine, torque is varied by inserting or withdrawing pairs of thin sluiceplates between rotor and stator, thus controlling the extent of the development of thetoroidal vortices.

Fixed

casing

Fixed

casing

Fixed casing

Rotor

Rotor

Rotation

Dynamometer shaft

a a

(a) (b)

(c) (d)

Figure 8.9 Hydrokinetic dynamometer, principle of operation: (a) sectionthrough dynamometer; (b) end view of rotor; (c) development of section a–aof rotor and casing; (d) representation of toroidal vortex

156 Engine Testing

Typical cross-section through casing of Froude dynamometer, type DPX

(1) Rotor

(2) Water outlet valve

(3) Water inlet valve

(4) Sluice plates for load control

(5) Water inlet holes in vanes

(6) Casing liners

(7) Casing trunnion bearing

(8) Shaft bearing

(9) Tachometer

4

5

6

7

8

10

1

2

3

9

Figure 8.10 Froude sluice-plate dynamometer

1(b) Variable fill machines, Fig. 8.11. In these machines, the torque absorbed isvaried by adjusting the quantity (mass) of water in circulation within the casing. Thisis achieved by a valve, usually on the water outlet, associated with control systemsof widely varying complexity. The particular advantage of the variable fill machine


Figure 8.11 Variable fill hydraulic dynamometer controlled by fast acting outletvalve at bottom of the stator

is that the torque may be varied much more rapidly than is the case with sluice platecontrol. Amongst this family of machines are the largest dynamometers ever madewith rotors of around 5m diameter. There are several designs of water control valveand valve actuating mechanisms depending on the range and magnitude of the loadsabsorbed and the speed of change of load required. For the fastest response, it isnecessary to have adequate water available to fill the casing rapidly and it may benecessary to fit both inlet and outlet control valves with an integrated control system.

1(c) ‘Bolt-on’ variable fill machines. These machines, available for many yearsin the USA, operate on the same principle as those described in 1(b) above, but arearranged to bolt directly on to the engine clutch housing or into the truck chassis (seeChapter 4 for more detailed description of use). Machines are available for ratings upto about 1000 kW. In these machines, load is usually controlled by an inlet controlvalve associated with a throttled outlet. By nature of their simplified design andlower mass, these machines are not capable of the same level of speed holding ortorque measurement as the more conventional 1(b) designs.

1(d) Disc dynamometers. These machines, not very widely used, consist of one ormore flat discs located between flat stator plates, with a fairly small clearance. Poweris absorbed by intensive shearing of the water and torque is controlled as in variable

158 Engine Testing

fill machines. Disc dynamometers have comparatively poor low speed performancebut may be built to run at very high speeds, making them suitable for loading gasturbines. A variation is the perforated disc machine, in which there are holes in therotor and stators, giving greater power dissipation for a given size of machine.

2. Hydrostatic dynamometers. Not very widely used, these machines consist gen-erally of a combination of a fixed stroke and a variable stroke positive displacementhydraulic pump/motor similar to that found in large off-road vehicle transmissions.The fixed stroke machine forms the dynamometer. An advantage of this arrangementis that, unlike most other, non-electrical machines, it is capable of developing fulltorque down to zero speed and is also capable of acting as a source of power to‘motor’ the engine under test.

3. Electrical motor-based dynamometers. The common feature of all thesemachines is that the power absorbed is transformed into electrical energy, whichis ‘exported’ from the machine via its associated ‘drive’ circuitry. The energy losswithin both the motor and its drive in the form of heat is transferred to a coolingmedium, which may be water or is more commonly forced air flow.

All motor-based dynamometers have associated with them large drive cabinetsthat produce heat and noise. The various sections of these cabinets contain highvoltage/power devices and complex electronics; they have to be housed in suitableconditions which have a clean and non-condensing atmosphere with sufficient spacefor access and cooling. When planning a facility layout, the designer should rememberthat these large and heavy cabinets have to be positioned after the building work hasbeen completed. The position of the drives should normally be within 15m of thedynamometer, but this should be minimized so far as is practical to reduce the highcost of the connecting power cables.

3(a) Direct current (d.c.) dynamometers. These machines consist essentially of atrunnion-mounted d.c. motor generator. Control is almost universally by means of athyristor based a.c.→d.c.→a.c. converter.

These machines have a long pedigree in the USA, are robust, easily controlled,and capable of motoring and starting as well as of absorbing power. Disadvantagesinclude limited maximum speed and high inertia, which can present problems oftorsional vibration (see Chapter 9) and limited rates of speed change. Because theycontain a commutator, the maintenance of d.c. machines may be higher than thosebased on a.c. squirrel cage motors.

3(b) Asynchronous or alternating current (a.c.) dynamometers. These asyn-chronous machines consist essentially of an induction motor with squirrel cage rotor,the speed of which is controlled by varying the supply frequency. The modern powercontrol stage of the control will invariably be based upon insulated gate bipolartransistor (IGBT) technology.

The squirrel cage rotor machines have a lower rotational inertia than d.c. machinesof the same power and are therefore capable of better transient performance. Beingbased on an asynchronous motor they have proved very robust in service requiringlow maintenance.


However, it is misleading to think that any motor’s mechanical design may be usedwithout adaptation as a dynamometer. During the first decade of their wide industrialuse, it was discovered that several different dynamometer/motor designs sufferedfrom bearing failures caused by an electrical arcing effect within the rolling elements;this was due to the fact that, in their dynamometer role, a potential differencedeveloped between the rotor and the stator (ground). Ceramic bearing elements andother design features are now used to prevent such damage occurring.

3(c) Synchronous, permanent magnet dynamometers. The units represent thelatest generation of dynamometer development and while using the same drivetechnology as the asynchronous dynamometers are capable of higher dynamic per-formance because of their inherently lower rotational inertia. It is this generation ofmachine that will provide the high dynamic test tools required by engine and vehiclesystem simulation in the test cell.

Acceleration rates of 160 000 rpm/s and air-gap torque rise times of less than1ms have been achieved, which makes it possible to use these machines as enginesimulators where the full dynamic fluctuation speed and torque characteristic of theengine is required for drive-line component testing.

3(d) Eddy-current dynamometers, Fig. 8.3. These machines make use of theprinciple of electromagnetic induction to develop torque and dissipate power.A toothed rotor of high-permeability steel rotates, with a fine clearance, betweenwater-cooled steel loss plates. A magnetic field parallel to the machine axis is gen-erated by two annular coils and motion of the rotor gives rise to changes in thedistribution of magnetic flux in the loss plates. This in turn gives rise to circulatingeddy currents and the dissipation of power in the form of electrical resistive losses.Energy is transferred in the form of heat to cooling water circulating through pas-sages in the loss plates, while some cooling is achieved by the radial flow of air inthe gaps between rotor and plates.

Power is controlled by varying the current supplied to the annular exciting coilsand rapid load changes are possible. Eddy-current machines are simple and robust,the control system is simple and they are capable of developing substantial brakingtorque at quite low speeds. Unlike a.c. or d.c. dynamometers, however, they areunable to develop motoring torque.

There are two common forms of machine both having air circulating in the gapbetween rotor and loss (cooling) plates, hence ‘dry gap’:

1. Dry gap machines fitted with one or more tooth disk rotors. These machineshave lower inertia than the drum machines and a very large installed user base,particularly in Europe. However, the inherent design features of their loss platesplace certain operational restrictions on their use. It is absolutely critical tomaintain the required water flow through the machines at all times; even a veryshort loss of cooling will cause the loss plates to distort leading to the rotor/plategap closing with disastrous results. These machines must be fitted with flowdetection devices interlocked with the cell control system; pressure switchesshould not be used since in a closed water system it is possible to have pressurewithout flow.

160 Engine Testing

2. Dry gap machines fitted with a drum rotor. These machines usually have a higherinertia than the equivalent disc machine, but may be less sensitive to coolingwater conditions.

Although no longer so widely used, an alternative form of eddy-current machine isalso available. This employs a simple disc or drum design of rotor in which eddycurrents are induced and the heat developed is transferred to water circulated throughthe gaps between rotor and stator. These ‘wet gap’ machines are liable to corrosionif left static for any length of time, have higher inertia, and have a high level ofminimum torque, arising from drag of the cooling water in the gap.

4. Friction dynamometers, Fig. 8.12. These machines, in direct line of successionfrom the original rope brake, consist essentially of water-cooled, multidisc frictionbrakes. They are useful for low-speed applications, for example for measuring thepower output of a large, off-road vehicle transmission at the wheels, and have theadvantage, shared with the hydrostatic dynamometer, of developing full torque downto zero speed.

5. Air brake dynamometers. These devices, of which the Walker fan brake wasthe best-known example, are now largely obsolete. They consisted of a simplearrangement of radially adjustable paddles that imposed a torque that could beapproximately estimated. They survive mainly for use in the field testing of helicopterengines, where high accuracy is not required and the noise is no disadvantage.

Hybrid and tandem dynamometers

For completeness, mention should be made of both a combined design that isoccasionally adopted for cost reasons and the use of two dynamometers in line forspecial test configurations.

Figure 8.12 Water-cooled friction brake used as a dynamometer


The d.c. or a.c. electrical dynamometer is capable of generating a motoring torquealmost equal to its braking torque. However, the motoring torque required in enginetesting seldom exceeds 30 per cent of the engine power output. Since, for equal powerabsorption, a.c. and d.c. machines are more expensive than other types, it is sometimesworth running an electrical dynamometer in tandem with, for example, a variable fillhydraulic machine. Control of these hybrid machines is a more complex matter andthe need to provide duplicate services, both electrical power and cooling water, is afurther disadvantage. The solution may, however, on occasion be cost-effective.

Tandem machines are used when the torque/speed envelope of the prime movercannot be covered by a standard dynamometer, usually this is found in gas turbinetesting when the rotational speed is too high for a machine fitted with a rotor capableof absorbing full rated torque. The first machine in line has to have a shaft systemcapable of transmitting the combine torques.

Tandem machines are also used when the prime mover is producing power throughtwo contrarotating shafts as with some aero and military applications; in these casesthe first machine in line is of a special design with a hollow rotor shaft to allow thehousing of a quill shaft connecting the second machine.

One, two or four quadrant?

Figure 8.8 illustrates diagrammatically the four ‘quadrants’ in which a dynamometermay be required to operate. Most engine testing takes place in the first quadrant, theengine running anticlockwise when viewed on the flywheel end. On occasions it isnecessary for a test installation using a unidirectional water brake to accept enginesrunning in either direction; one solution is to fit the dynamometer with couplings atboth ends mounted on a turntable. Large and some ‘medium speed’ marine enginesare usually reversible.

All types of dynamometer are naturally able to run in the first (or second) quadrant.Hydraulic dynamometers are usually designed for one direction of rotation, thoughthey may be run in reverse at low fill state without damage. When designed specif-ically for bidirectional rotation they may be larger than a single-direction machineof equivalent power and torque control may not be as precise as that of the uni-directional designs. The torque measuring system must of course operate in bothdirections. Eddy-current machines are inherently reversible.

When it is required to operate in the third and fourth quadrants (i.e. for thedynamometer to produce power as well as to absorb it) the choice is effectivelylimited to d.c. or a.c. machines, or to the hydrostatic or hybrid machine. Thesemachines are generally reversible and therefore operate in all four quadrants.

There is an increasing requirement for four-quadrant operation as a result of thegrowth in transient testing, with its call for very rapid load changes and even fortorque reversals.

If mechanical losses in the engine are to be measured by ‘motoring’, a four-quadrant machine is obviously required.

162 Engine Testing

Table 8.2 Operating quadrants of dynamometer designs

Type of machine Quadrant

Hydraulic sluice plate 1 or 2Variable fill hydraulic 1 or 2‘Bolt on’ variable fill hydraulic 1 or 2Disc type hydraulic 1 and 2Hydrostatic 1, 2, 3, 4d.c. electrical 1, 2, 3, 4a.c. electrical 1, 2, 3, 4Eddy current 1 and 2Friction brake 1 and 2Air brake 1 and 2Hybrid 1, 2, 3, 4

A useful feature of such a machine is its ability also to start the engine. Table 8.2summarizes the performance of machines in this respect.

Matching engine and dynamometer characteristics

The different types of dynamometer have significantly different torque–speed andpower–speed curves, and this can affect the choice made for a given application.Figure 8.13 shows the performance curves of a typical hydraulic dynamometer. Thedifferent elements of the performance envelope are as follows:

• Dynamometer full (or sluice plates wide open). Torque increases with square ofspeed, no torque at rest.

• Performance limited by maximum permitted shaft torque.• Performance limited by maximum permitted power, which is a function of cooling

water throughput and its maximum permitted temperature rise.• Maximum permitted speed.• Minimum torque corresponding to minimum permitted water flow.

Figure 8.14 shows the considerably different performance envelope of an electricalmachine, made up of the following elements:

• Constant torque corresponding to maximum current and excitation.• Performance limited by maximum permitted power output of machine.• Maximum permitted speed.


100%

100% 100%

100%

Torq

ue

Pow

er

Speed Speed

c

cb

d

e

da

a

b

e

0 0

Figure 8.13 Engine torque curves plotted on hydraulic dynamometer torquecurves

100%

100% 100%

100%

Speed Speed0 0

a

b

c

a

b

c

Torq

ue

Pow

er

Figure 8.14 Performance curve shapes of electrical motor-baseddynamometers

Since these are ‘four-quadrant’ machines, power absorbed can be reduced to zeroand there is no minimum torque curve.

Figure 8.15 shows the performance curves for an eddy-current machine, whichlie between those of the previous two machines:

• Low speed torque corresponding to maximum permitted excitation.• Performance limited by maximum permitted shaft torque.• Performance limited by maximum permitted power, which is a function of cooling

water throughput and its maximum permitted temperature rise.• Maximum permitted speed.• Minimum torque corresponding to residual magnetization, windage and friction.

In choosing a dynamometer for an engine or range of engines, it is essential tosuperimpose the maximum torque– and power–speed curves on to the dynamometer

164 Engine Testing

100%

100% 100%

100%

Speed Speed0 0

a

d

d

e e

b

c

a

b

c

Torq

ue

Pow

er

Figure 8.15 Performance curve shapes of an eddy-current dynamometer

envelope. See the example in Fig. 8.13 which demonstrates a typical problem: thehydraulic machine is incapable of developing sufficient torque at the bottom end ofthe speed range.

For best accuracy, it is desirable to choose the smallest machine that will copewith the largest engine to be tested. Hydraulic dynamometers are generally able todeal with a moderate degree of overload and overspeed, but it is undesirable to runelectrical machines beyond their rated limits: this can lead to damage to commutators,overheating and distortion of eddy-current loss plates.

Careful attention must also be given to the arrangements for coupling engine anddynamometer, see Chapter 9.

Engine starting and cranking

Starting an engine when it is connected to a dynamometer may present the celldesigner and operator with a number of problems, and is a factor to be borne inmind when selecting the dynamometer. If the engine is fitted with a starter motor,the cell system must provide the high current d.c. supply and associated switching;in the absence of an engine mounted starter a complete system to start and crankthe engine must be available which compromises neither the torsional characteristics(see Chapter 9) nor the torque measurement accuracy.

Engine cranking, no starter motor

The cell cranking system must be capable of accelerating the engine to its normalstarting speed and, in most cases, of disengaging when the engine fires. A four-quadrant dynamometer, suitably controlled, will be capable of starting the enginedirectly. The power available from any four-quadrant machine will always be greaterthan that required, therefore excessive starting torque must be avoided by an alarm


system otherwise an engine locked by seizure or fluid in a cylinder may cause damageto the drive line.

The preferred method of providing other types of dynamometer with a startingsystem is to mount an electric motor at the non-engine end of the dynamometer shaft,driving through an over-running or remotely engaged clutch, and generally througha speed-reducing belt drive. The clutch half containing the mechanism should beon the input side, otherwise it will be affected by the torsional vibrations usuallyexperienced by dynamometer shafts. The motor may be mounted above, below oralongside the dynamometer to save cell length.

The sizing of the motor must take into account the maximum break-away torqueexpected, usually estimated as twice the average cranking torque, while the normalrunning speed of the motor should correspond to the desired cranking speed. Thechoice of motor and associated starter must take into account the maximum numberof starts per hour that may be required, both in normal use and when dealing witha faulty engine. The running regime of the motor is demanding, involving repeatedbursts at overload, with the intervening time at rest, and an independent cooling fanmay be necessary.

Some modern diesel engines, when ‘green’,∗ require cranking at more than the nor-mal starting speed, sometimes as high as 1200 rev/min, in order to prime the fuel sys-tem. In such cases a two-speed or fully variable speed starter motor may be necessary.

The system must be designed to impose the minimum parasitic torque when dis-engaged, since this torque will not be sensed by the dynamometer measuring system.

In some cases, to avoid this source of inaccuracy, the motor may be mounteddirectly on the dynamometer carcase and permanently coupled to the dynamometershaft by a belt drive. This imposes an additional load on the trunnion bearings, whichmay lead to brinelling, and it also increases the effective moment of inertia of thedynamometer. However, it has the advantage that motoring and starting torque maybe measured by the dynamometer system.

An alternative solution is to use a standard vehicle engine starter motor in con-junction with a gear ring carried by a ‘dummy flywheel’ carried on a shaft withseparate bearings incorporated in the drive line, but this may have the disadvantageof complicating the torsional behaviour of the system.

Engine-mounted starter systems

If the engine is fitted with its own starter motor on arrival at the test stand, all thatmust be provided is the necessary 12 or 24V supply. The traditional approach hasbeen to locate a suitable battery as close as possible to the starter motor, with asuitable battery charger supply. This system is not ideal, as the battery needs to be

∗ A green engine is one that has never been run. The rubbing surfaces may be dry, the fuelsystem may need priming, and there is always the possibility that it, or its control system,is faulty and incapable of starting.

166 Engine Testing

in a suitably ventilated box, to avoid the risk of accidental shorting, and will take upvaluable space. Special transformer/rectifier units designed to replace batteries forthis duty are on the market. They will include an ‘electrical services box’ to providepower in addition for ignition systems and diesel glow plugs. In large integratedsystems there may be a bus bar system for the d.c. supplies.

The engine starter will be presented with a situation not encountered in normalservice: it will be required to accelerate the whole dynamometer system in additionto the engine while a ‘green’ engine may exhibit a very high breakaway torque andrequire prolonged cranking at high speed to prime the fuel system before it fires.

Non-electrical starting systems

Diesel engines larger than the automotive range are usually started by means of com-pressed air, admitted to the cylinders by way of starting valves. In some cases it isnecessary to move the crankshaft to the correct starting position, either by barring orusing an engine-mounted inching motor. The test facility should include a compressorand a receiver of capacity at least as large as that recommended for the engine in service.

Compressed air or hydraulic motors are sometimes used instead of electric motorsto provide cranking power but have no obvious advantages over a d.c. electric motor,apart from a marginally reduced fire risk in the case of compressed air, provided thesupply is shut off automatically in the case of fire.

In Chapter 9, attention is drawn to the possibility of overloading flexible couplingsin the drive line during the starting process, and particularly when the engine firstfires. This should not be overlooked.

Choice of dynamometer

Table 8.3 lists the various types of dynamometer and indicates their applicability forvarious classes of engine being tested in steady or mild transient states.

In most cases, several choices are available and it will be necessary to consider thespecial features of each type of dynamometer and to evaluate the relative importanceof these in the particular case. These features are listed in Table 8.3 and other specialfactors are considered later.

Some additional considerations

The final choice of dynamometer for a given application may be influenced by someof the following factors:

1. The speed of response required by the test sequences being run: steady state,transient, dynamic or high dynamic. This will determine the technology andprobably the number of quadrants of operation required.


Table 8.3 Dynamometers: advantages and disadvantage of available types

Dynamometer type Advantages Disadvantages

Froude sluice plate Obsolete, but manycheap andreconditioned modelsin use worldwide,robust

Slow response to change inload. Manual control noteasy to automate

Variable fill waterbrakes

Capable of mediumspeed load change,automated control,robust and tolerant ofoverload. Availablefor largestprime-movers

‘Open’ water systemrequired. Can suffer fromcavitation or corrosiondamage

‘Bolt-on’ variablefill water brakes

Cheap and simpleinstallation. Up to1000 kW

Lower accuracy ofmeasurement and controlthan fixed machines

Disc type hydraulic Suitable for high speeds Poor low speedperformance

Hydrostatic For special applications,provides fourquadrant performance

Mechanically complex,noisy and expensive.System contains largevolumes of high pressureoil

d.c electrical motor Mature technology.Four quadrantperformance

High inertia, commutatormay be fire andmaintenance risk

asychronous motor(a.c.)

Lower inertia than d.c.Four quadrantperformance

Expensive. Large drivecabinet needs suitablehousing

Permanent magnetmotor

Lowest inertia,most dynamicfour quadrantperformance. Smallsize in cell

Expensive. Large drivecabinet needs suitablehousing

Eddy current Low inertia (disk typeair gap). Well adaptedto computer control.Mechanically simple

Vulnerable to poor coolingsupply. Not suitable forsustained rapid changesin power (thermalcycling)

(continued)

168 Engine Testing

Table 8.3 (cont.)

Dynamometer type Advantages Disadvantages

Friction brake Special purposeapplications for veryhigh torques at lowspeed

Limited speed range

Air brake Cheap. Very littlesupport servicesneeded

Noisy. Limited controlaccuracy

Hybrid Possible cost advantageover sole electricalmachine

Complexity of constructionand control

2. Load factor. If the machine is to spend long periods out of use, the possibilitiesof corrosion must be considered, particularly in the case of hydraulic or wetgap eddy-current machines. Can the machine be drained readily? Should theuse of corrosion inhibitors be considered?

3. Overloads. If it may be necessary to consider occasional overloading of themachine a hydraulic machine may be preferable, in view of its greater toler-ance of such conditions. Check that the torque measuring system has adequatecapacity.

4. Large and frequent changes in load. This can give rise to problems with eddy-current machines, due to expansion and contraction with possible distortion ofthe loss plates.

5. Wide range of engine sizes to be tested. It may be difficult to achieve goodcontrol and adequate accuracy when testing the smallest engines, while theminimum dynamometer torque may also be inconveniently high.

6. How are engines to be started? If a non-motoring dynamometer is favouredit may be necessary to fit a separate starter to the dynamometer shaft. Thisrepresents an additional maintenance commitment and may increase inertia.

7. Is there an adequate supply of cooling water of satisfactory quality? Hard waterwill result in blocked cooling passages and some water treatments can give riseto corrosion. This may be a good reason for choosing d.c. or a.c. dynamometers,despite extra cost.

8. Is the pressure of the water supply subject to sudden variations? Sudden pressurechanges or regular pulsations will affect the stability of control of hydraulicdynamometers. Eddy-current and indirectly cooled machines are unaffectedproviding inlet flow does not fall below emergency trip levels.

9. Is the electrical supply voltage liable to vary as the result of other loads onthe same circuit? With the exception of air brakes and manually controlledhydraulic machines, all dynamometers are affected by electrical interferenceand voltage changes.


10. Is it proposed to use a shaft docking system for coupling engine and dynamome-ter? Are there any special features or heavy overhung or axial loads associ-ated with the coupling system? Such features should be discussed with thedynamometer manufacturer before making a decision. Some machines, notablythe Schenck flexure plate mounting system, are not suited to taking axial loads.

The supplier of any new dynamometer should offer an acceptance test and basictraining in operation, calibration and safety of the new machine. A careful checkshould be made on the level of technical support, including availability of calibrationservices, spares and local service facilities, offered by the manufacturer.

9 Coupling the engine to thedynamometer

Introduction

The selection of suitable couplings and shaft for the connection of the engine tothe dynamometer is by no means a simple matter. Incorrect choice or faulty systemdesign may give rise to a number of problems:

• torsional oscillations;• vibration of engine or dynamometer;• whirling of coupling shaft;• damage to engine or dynamometer bearings;• excessive wear of shaft line components;• catastrophic failure of coupling shafts;• engine starting problems.

This whole subject, the coupling of engine and dynamometer, can give rise to moretrouble than any other routine aspect of engine testing, and a clear understanding ofthe many factors involved is desirable. The following chapter covers all the mainfactors, but in some areas more extensive analysis may be necessary and appropriatereferences are given.

The nature of the problem

The special feature of the problem is that it must be considered afresh each timean engine not previously encountered is installed. It must also be recognized thatunsatisfactory torsional behaviour is associated with the whole system – engine,coupling shaft and dynamometer – rather than with the individual components, allof which may be quite satisfactory in themselves.

Problems arise partly because the dynamometer is seldom equivalent dynamicallyto the system driven by the engine in service. This is particularly the case withvehicle engines. In the case of a vehicle with rear axle drive, the driveline consistsof a clutch, which may itself act as a torsional damper, followed by a gearbox, thecharacteristics of which are low inertia and some damping capacity. This is followed

Coupling the engine to the dynamometer 171

Figure 9.1 Simple form of dynamometer/engine drive line

by a drive shaft and differential, itself having appreciable damping, two half shaftsand two wheels, both with substantial damping capacity and running at much slowerspeed than the engine, thus reducing their effective inertia.

When coupled to a dynamometer this system, Fig. 9.1, with its built-in dampingand moderate inertia, is replaced by a single drive shaft connected to a single rotatingmass, the dynamometer, running at the same speed as the engine. The clutch may ormay not be retained.

Particular care is necessary where the moment of inertia of the dynamometeris more than about twice that of the engine. A further consideration that must betaken seriously concerns the effect of the difference between the engine mountingarrangements in the vehicle and on the testbed. This can lead to vibrations of thewhole engine that can have a disastrous effect on the drive shaft.

Overhung mass on engine and dynamometer bearings

Care must be taken when designing and assembling a shaft system that the loadsimposed by the mass and unbalanced forces do not exceed the overhung weightlimits of the engine bearing at one end and the dynamometer at the other. Steeladaptor plates required to adapt the bolt holes of the shaft to the dynamometer flangeor engine flywheel can increase the load on bearings significantly and compromisethe operation of the system. Dynamometer manufacturers produce tables showingthe maximum permissible mass at a given distance from the coupling face of theirmachines; the equivalent details for most engines is more difficult to obtain, but thedanger of overload should be kept in mind by all concerned.

Background reading

The mathematics of the subject is complex and not readily accessible. Den Hartog1

gives what is possibly the clearest exposition of fundamentals. Ker Wilson’s clas-sical treatment in five volumes2 is probably still the best source of comprehensiveinformation; his abbreviated version3 is sufficient for most purposes. Mechanical

172 Engine Testing

Engineering Publications have published a useful practical handbook4 while Lloyd’sRegister5 gives rules for the design of marine drives that are also useful in the presentcontext. A listing of the notation used is to be found at the end of this chapter.

Torsional oscillations and critical speeds

In its simplest form, the engine–dynamometer system may be regarded as equivalentto two rotating masses connected by a flexible shaft, Fig. 9.2. Such a system hasan inherent tendency to develop torsional oscillations. The two masses can vibrate180� out of phase about a node located at some point along the shaft between them.The oscillatory movement is superimposed on any steady rotation of the shaft. Theresonant or critical frequency of torsional oscillation of this system is given by:

nc =60

2�

√

Cc�Ie+ Ib�

IeIb(1)

If an undamped system of this kind is subjected to an exciting torque of constantamplitude Tex and frequency n, the relation between the amplitude of the inducedoscillation � and the ratio n/nc is as shown in Fig. 9.3.

At low frequencies, the combined amplitude of the two masses is equal to thestatic deflection of the shaft under the influence of the exciting torque, �0 = Tex/Cs.As the frequency increases, the amplitude rises and at n= nc it becomes theoreticallyinfinite: the shaft may fracture or non-linearities and internal damping may preventactual failure. With further increases in frequency the amplitude falls and at n−

√

2nc

it is down to the level of the static deflection. Amplitude continues to fall withincreasing frequency.

The shaft connecting engine and dynamometer must be designed with a suitablestiffness Cs to ensure that the critical frequency lies outside the normal operating

le

StiffnessC

l b

Figure 9.2 Two mass system (compare with Fig. 9.1)


Mc = 5

Undamped

Damped

M =

θ/θ

0

0 1

n/nc

1.5 20.5

5

4

3

2

0

θ0 = 1

Figure 9.3 Relationship between frequency ratio, amplitude and dynamicamplifier M

range of the engine, and also with a suitable degree of damping to ensure that the unitmay be run through the critical speed without the development of a dangerous levelof torsional oscillation. Figure 9.3 also shows the behaviour of a damped system.The ratio �/�0 is known as the dynamic magnifier M. Of particular importance is thevalue of the dynamic magnifier at the critical frequency, Mc. The curve of Fig. 9.3corresponds to a value Mc = 5.

Torsional oscillations are excited by the variations in engine torque associatedwith the pressure cycles in the individual cylinders (also, though usually of lessimportance, by the variations associated with the movement of the reciprocatingcomponents).

Figure 9.4 shows the variation in the case of a typical single cylinder four-strokediesel engine. It is well known that any periodic curve of this kind may be synthesizedfrom a series of harmonic components, each a pure sine wave of a different amplitudehaving a frequency corresponding to a multiple or submultiple of the engine speedand Fig. 9.4 shows the first six components.

The order of the harmonic defines this multiple. Thus a component of orderN0 = 1/2 occupies two revolutions of the engine, N0 = 1 one revolution and so on.In the case of a four cylinder four-stroke engine, there are two firing strokes perrevolution of the crankshaft and the turning moment curve of Fig. 9.4 is repeatedat intervals of 180�. In a multicylinder engine, the harmonic components of a given

174 Engine Testing

Torque

Crank angle

Mmean

0 180° 360° 540° 720°

½-order

1-order

2-order

3-order

2½-order

1½-order

Figure 9.4 Harmonic components of turning moment, single cylinder four-stoke gasoline engine

order for the individual cylinders are combined by representing each component bya vector, in the manner illustrated in Chapter 3, for the inertia forces. A completetreatment of this process is beyond the scope of this book, but the most significantresults may be summarized as follows.

The first major critical speed for a multicylinder, in-line engine is of order:

N0 = NCYL/2 for a four-stroke engine (2a)


N0 = NCYL for a two-stroke engine (2b)

Thus, in the case of a four cylinder, four-stroke engine the major critical speeds areof order 2, 4, 6, etc. In the case of a six cylinder engine, they are of order 3, 6, 9, etc.

The distinction between a major and a minor critical speed is that in the case of anengine having an infinitely rigid crankshaft it is only at the major critical speeds thattorsional oscillations can be induced. This, however, by no means implies that in largeengines having a large number of cylinders, the minor critical speeds may be ignored.

At the major critical speeds the exciting torques Tex of all the individual cylindersin one line act in phase and are thus additive (special rules apply governing thecalculation of the combined excitation torques for Vee engines).

The first harmonic is generally of most significance in the excitation of torsionaloscillations, and for engines of moderate size, such as passenger vehicle engines, it isgenerally sufficient to calculate the critical frequency from eq. (1), then to calculatethe corresponding engine speed from:

Nc = nc/N0 (3)

The stiffness of the connecting shaft between engine and dynamometer should bechosen so that this speed does not lie within the range in which the engine is requiredto develop power.

In the case of large multicylinder engines, the ‘wind-up’ of the crankshaft as aresult of torsional oscillations can be very significant and the two-mass approximationis inadequate; in particular, the critical speed may be overestimated by as much as20 per cent and more elaborate analysis is necessary. The subject is dealt with inseveral different ways in the literature; perhaps the easiest to follow is that of DenHartog.1 The starting point is the value of the mean turning moment developed bythe cylinder, Mmean (Fig. 9.4). Values are given for a so-called ‘p factor’, by whichMmean is multiplied to give the amplitude of the various harmonic excitation forces.Table 9.1, reproduced from Den Hartog, shows typical figures for a four-strokemedium speed diesel engine.

Exciting torque:

Tex = p� Mmean (4)

Table 9.1 p factors

Order 12 1 1 1

2 2 2 12 3� � � 8

p factor 2.16 2.32 2.23 1.91 1.57 1.28 0.08

176 Engine Testing

The relation between Mmean and imep (indicated mean effective pressure) isgiven by:

for a four-stroke engine Mmean = pi�B2S

16�10−4 (5a)

for a two-stroke engine Mmean = pi�B2S

8�10−4 (5b)

Lloyd’s Rulebook,5 the main source of data on this subject, expresses the amplitudeof the harmonic components rather differently, in terms of a ‘component of tangentialeffort’, Tm. This is a pressure that is assumed to act upon the piston at the crankradius S/2. Then exciting torque per cylinder:

Tex = Tm

�B2

4

S

2×10−4 (6)

Lloyd’s give curves of Tm in terms of the indicated mean effective pressure pi andit may be shown that the values so obtained agree closely with those derived fromTable 9.1.

The amplitude of the vibratory torque Tv induced in the connecting shaft by thevector sum of the exciting torques for all the cylinders,

∑

Tex is given by:

Tv =∑

TexMc

�1+ Ie/Ib�(7)

The complete analysis of the torsional behaviour of a multicylinder engine is asubstantial task, though computer programmes are available which reduce the effortrequired. As a typical example, Fig. 9.5 shows the ‘normal elastic curves’ for the firstand second modes of torsional oscillation of a 16 cylinder Vee engine coupled to ahydraulic dynamometer. These curves show the amplitude of the torsional oscilla-tions of the various components, relative to that at the dynamometer which is taken asunity. The natural frequencies are respectively nc = 4820 c.p.m. and nc = 6320 c.p.m.The curves form the basis for further calculations of the energy input giving rise to theoscillation. In the case of the engine under consideration, these showed a very severefourth-order oscillation, N0 = 4, in the first mode. (For an engine having eight cylin-ders in line the first major critical speed, from eq. (2a), is of order N0 = 4.) The enginespeed corresponding to the critical frequency of torsional oscillation is given by:

Nc = nc/N0 (8)

giving, in the present case, Nc = 1205 rev/min, well within the operating speedrange of the engine. Further calculations showed a large input of oscillatory energyat N0 = 4½, a minor critical speed, in the second mode, corresponding to a criticalengine speed of 6320/4¼= 1404 rev/min, again within the operating range. Several


Second mode 6320 c.p.m.

First mode 4820 c.p.m.

Cylinder

Dam

per

Dam

per

hub

Dynam

om

ete

r

Fly

wheel

Rela

tive a

mplit

ude

Fla

nge

1

4

3

2

1

0

–1

–2

–3

–4

2 3 4 5 6 7 8

Figure 9.5 Normal elastic curves for a particular 16 cylinder V engine coupledto a hydraulic dynamometer (taken from an actual investigation)

failures of the shaft connecting engine and dynamometer occurred before a safesolution was arrived at.

This example illustrates the need for caution and for full investigation in setting

up large engines on the test bed.

It is not always possible to avoid running close to or at critical speeds and thissituation is usually dealt with by the provision of torsional vibration dampers, in whichthe energy fed into the system by the exciting forces is absorbed by viscous shearing.Such dampers are commonly fitted at the non-flywheel end of engine crankshafts. Insome cases it may also be necessary to consider their use as a component of enginetest cell drive lines, when they are located either as close as possible to the engineflywheel, or at the dynamometer. The damper must be ‘tuned’ to be most effectiveat the critical frequency and the selection of a suitable damper involves equating the

178 Engine Testing

energy fed into the system per cycle with the energy absorbed by viscous shear inthe damper. This leads to an estimate of the magnitude of the oscillatory stresses atthe critical speed. For a clear treatment of the theory, see Den Hartog.1

Points to remember:

• As a general rule, it is good practice to avoid running the engine under powerat speeds between 0.8 and 1.2 times critical speed. If it is necessary to take theengine through the critical speed, this should be done off load and as quickly aspossible. With high inertia dynamometers the transient vibratory torque may wellexceed the mechanical capacity of the drive line and the margin of safety of thedrive line components may need to be increased.

• Problems frequently arise when the inertia of the dynamometer much exceeds thatof the engine: a detailed torsional analysis is desirable when this factor exceeds2. This situation usually occurs when it is found necessary to run an engine ofmuch smaller output than the rated capacity of the dynamometer.

• the simple ‘two mass’ approximation of the engine-dynamometer system is inad-equate for large engines and may lead to overestimation of the critical speed.

Design of coupling shafts

The maximum shear stress induced in a shaft, diameter D, by a torque T Nm isgiven by:

=16T

�D3Pa (9a)

In the case of a tubular shaft, bore diameter d, this becomes:

=16TD

��D4−d4�Pa (9b)

For steels, the shear yield stress is usually taken as equal to 0.75× yield stress intension. A typical choice of material would be a nickel–chromium–molybdenumsteel, to specification BS 817M40 (previously En 24) heat-treated to the ‘T’ condition.

The various stress levels for this steel are roughly as follows:

ultimate tensile strength not less than 850MPa (55 t.s.i.)0.1% proof stress in tension 550MPaultimate shear strength 500MPa0.1% proof stress in shear 300MPashear fatigue limit in reversed stress ±200MPa

It is clear that the permissible level of stress in the shaft will be a small fraction ofthe ultimate tensile strength of the material.


The choice of designed stress level at the maximum rated steady torque is influ-enced by two principal factors.

Stress concentrations, keyways and keyless hub connection

For a full treatment of the very important subject of stress concentration see Ref. 6.There are two particularly critical locations to be considered:

• At a shoulder fillet, such as is provided at the junction with an integral flange. Fora ratio fillet radius/shaft diameter= 0.1 the stress concentration factor is about1.4, falling to 1.2 for r/d= 0.2.

• At the semicircular end of a typical rectangular keyway, the stress concentrationfactor reaches a maximum of about 2.5× nominal shear stress at an angle ofabout 50� from the shaft axis. The authors have seen a number of failures at thislocation and angle.

Cyclic stresses associated with torsional oscillations is an important consideration andas, even in the most carefully designed installation involving an internal combustionengine, some torsional oscillation will be present, it is wise to select a conservativevalue for the nominal (steady state) shear stress in the shaft.

In view of the stress concentration inherent in shaft keyways and the backlashpresent that can develop in splined hubs, the use of keyless hub connection systemsof the type produced by the Ringfeder Corporation are now widely used. Thesedevices are supported by comprehensive design documentation; however, the actualinstallation process must be exactly followed for the design performance to beensured.

Stress concentration factors apply to the cyclic stresses rather than to the steadystate stresses. Figure 9.6 shows diagrammatically the Goodman diagram for a steelhaving the properties specified above. This diagram indicates approximately the

200

100

0

Mean shear stress (MPa)

Altern

ating s

tress (

MP

a)

0 100 200 300 400 500

Figure 9.6 Goodman diagram, steel shaft in shear

180 Engine Testing

relation between the steady shear stress and the permissible oscillatory stress. Theexample shown indicates that, for a steady torsional stress of 200MPa, the accom-panying oscillatory stress (calculated after taking into account any stress concentra-tion factors) should not exceed ±120MPa. In the absence of detailed design data,it is good practice to design shafts for use in engine test beds very conservatively,since the consequences of shaft failure can be so serious. A shear stress calculated inaccordance with eq. (9) of about 100MPa for a steel with the properties listed shouldbe safe under all but the most unfavourable conditions. To put this in perspective, ashaft 100mm diameter designed on this basis would imply a torque of 19 600Nm,or a power of 3100 kW at 1500 rev/min.

The torsional stiffness of a solid shaft of diameter D and length L is given by:

Cs =�D4G

32L(10a)

for a tubular shaft, bore d:

Cs =��D4−d4�

32L(10b)

Shaft whirl

The coupling shaft is usually supported at each end by a universal joint or flexiblecoupling. Such a shaft will ‘whirl’ at a rotational speed Nw (also at certain higherspeeds in the ratio 22 Nw, 3

2 Nw, etc.).The whirling speed of a solid shaft of length L is given by:

Nw =30�

L2

√

E�D4

64Ws

(11)

It is desirable to limit the maximum engine speed to about 0.8Nw.When using rubber

flexible couplings it is essential to allow for the radial flexibility of these couplings,

since this can drastically reduce the whirling speed. It is sometimes the practice tofit self-aligning rigid steady bearings at the centre of flexible couplings in high-speedapplications, but these are liable to give fretting problems and are not universallyfavoured.

As is well known, the whirling speed of a shaft is identical with its naturalfrequency of transverse oscillation. To allow for the effect of transverse couplingflexibility the simplest procedure is to calculate the transverse critical frequency ofthe shaft plus two half couplings from the equation:

Nt =30

�

√

k

W(12a)


where W=mass of shaft+ half couplings and k= combined radial stiffness of thetwo couplings.

Then whirling speed N taking this effect into account will be given by:

(

1

N

)2

=(

1

Nw

)2

+(

1

Nt

)2

(12b)

Couplings

The choice of the appropriate coupling for a given application is not easy: themajority of drive line problems probably have their origin in an incorrect choiceof components for the drive line, and are often cured by changes in this region.A complete discussion would much exceed the scope of this book, but the readerconcerned with drive line design should obtain a copy of Ref. 4, which gives acomprehensive treatment together with a valuable procedure for selecting the besttype of coupling for a given application. A very brief summary of the main types ofcoupling follows.

Quill shaft with integral flanges and rigid couplings

This type of connection is best suited to the situation where a driven machine ispermanently coupled to the source of power, when it can prove to be a simple andreliable solution. It is not well suited to test bed use, since it is intolerant of relativevibration and misalignment.

Quill shaft with toothed or gear type couplings

Gear couplings are very suitable for high powers and speeds, and can deal withrelative vibration and some degree of misalignment, but this must be very carefullycontrolled to avoid problems of wear and lubrication. Lubrication is particularlyimportant as once local tooth to tooth seizure takes place deterioration may be rapidand catastrophic. Such shafts are inherently stiff in torsion.

Conventional ‘cardan shaft’ with universal joints

These shafts are readily available from a number of suppliers, and are probablythe preferred solution in the majority of cases. However, standard automotive typeshafts can give trouble when run at speeds in excess of those encountered in vehicleapplications. A correct ‘built-in’ degree of misalignment is necessary to avoid frettingof the needle rollers.

182 Engine Testing

Figure 9.7 Multiple steel disc type flexible coupling

Multiple membrane couplings

These couplings, Fig. 9.7, are stiff in torsion but tolerant of a moderate degree ofmisalignment and relative axial displacement. They can be used for very high speeds.

Elastomeric element couplings

There is a vast number of different designs on the market and selection is not easy.Ref. 8 may be helpful. The great advantage of these couplings is that their torsionalstiffness may be varied widely by changing the elastic elements and problems asso-ciated with torsional vibrations and critical speeds dealt with (see the next section).

Damping: the role of the flexible coupling

The earlier discussion leads to two main conclusions: the engine–dynamometer sys-tem is susceptible to torsional oscillations and the internal combustion engine is apowerful source of forces calculated to excite such oscillations. The magnitude ofthese undesirable disturbances in any given system is a function of the dampingcapacities of the various elements: the shaft, the couplings, the dynamometer and theengine itself.

The couplings are the only element of the system the damping capacity of whichmay readily be changed, and in many cases, for example with engines of automotivesize, the damping capacity of the remainder of the system may be neglected, at leastin an elementary treatment of the problem, such as will be given here.

The dynamic magnifier M (Fig. 9.3) has already been mentioned as a measure ofthe susceptibility of the engine–dynamometer system to torsional oscillation. Nowreferring to Fig. 9.1, let us assume that there are two identical flexible couplings,of stiffness Cc, one at each end of the shaft, and that these are the only sources ofdamping. Figure 9.8 shows a typical torsionally resilient coupling in which torque istransmitted by way of a number of shaped rubber blocks or bushes which providetorsional flexibility, damping and a capacity to take up misalignment. The torsionalcharacteristics of such a coupling are shown in Fig. 9.9.


Figure 9.8 Rubber bush type torsionally resilient coupling

Steel shaft

–∆θ +∆θ

+∆T

–∆T

Torq

ue T

(N

m)

Angular deflection (θ)

a

b

T

Figure 9.9 Dynamic torsional characteristic of multiple bush type coupling

These differ in three important respects from those of, say, a steel shaft in torsion:

1. The coupling does not obey Hooke’s law: the stiffness or coupling rateCc =T/� increases with torque. This is partly an inherent property of therubber and partly a consequence of the way it is constrained.

2. The shape of the torque–deflection curve is not independent of frequency.Dynamic torsional characteristics are usually given for a cyclic frequency of10Hz. If the load is applied slowly the stiffness is found to be substantially less.The following values of the ratio dynamic stiffness (at 10Hz) to static stiffnessof natural rubber of varying hardness are taken from Ref. 4.

184 Engine Testing

Shore (IHRD) hardness 40 50 60 70Dynamic stiffnessStatic stiffness 1.5 1.8 2.1 2.4

Since the value of Cc varies with the deflection, manufacturers usually quote asingle figure which corresponds to the slope of the tangent ab to the torque–deflection curve at the rated torque, typically one third of the maximum permittedtorque.

3. If a cyclic torque ±T, such as that corresponding to a torsional vibration, issuperimposed on a steady torque T, Fig. 9.9, the deflection follows a path similarto that shown dotted. It is this feature, the hysteresis loop, which results in thedissipation of energy, by an amount W proportional to the area of the loop thatis responsible for the damping characteristics of the coupling.

Damping energy dissipated in this way appears as heat in the rubber and can, under

adverse circumstances, lead to overheating and rapid destruction of the elements.

The appearance of rubber dust inside coupling guards is a warning sign.

The damping capacity of a component such as a rubber coupling is described bythe damping energy ratio:

� =W

W

This may be regarded as the ratio of the energy dissipated by hysteresis in a singlecycle to the elastic energy corresponding to the wind-up of the coupling at meandeflection:

W =1

2T� =

1

2T 2/Cc

The damping energy ratio is a property of the rubber. Some typical values are givenin Table 9.2. The dynamic magnifier is a function of the damping energy ratio:as would be expected a high damping energy ratio corresponds to a low dynamicmagnifier. Some authorities give the relation:

M = 2�/�

Table 9.2 Damping energy ratio �

Shore (IHRD) hardness 50/55 60/65 70/75 75/80Natural rubber 0.45 0.52 0.70 0.90Neoprene 0.79Styrene-butadiene (SBR) 0.90


Table 9.3 Dynamic magnifier M

Shore (IHRD) hardness 50/55 60/65 70/75 75/80Natural rubber 10.5 8.6 5.2 2.7Neoprene 4.0Styrene-butadiene (SBR) 2.7

However, it is pointed out in Ref. 2 that for damping energy ratios typical of rubberthe exact relation:

� =(

1− e−2�/M)

is preferable. This leads to values of M shown in Table 9.3, which correspond to thevalues of � given in Table 9.2.

It should be noted that when several components, e.g. two identical rubber cou-plings, are used in series the dynamic magnifier of the combination is given by:

(

1

M

)2

=(

1

M1

)2

+(

1

M2

)2

+(

1

M3

)2

+· · · (13)

(this is an empirical rule, recommended in Ref. 5).

An example of drive shaft design

The application of these principles is best illustrated by a worked example. Figure 9.1represents an engine coupled by way of twin multiple-bush type rubber couplingsand an intermediate steel shaft to an eddy current dynamometer, with dynamometerstarting.

Engine specification is as follows:

Four cylinder four-stroke gasoline engineSwept volume 2.0 litre, bore 86mm, stroke 86mmMaximum torque 110Nm at 4000 rev/minMaximum speed 6000 rev/minMaximum power output 65 kWMaximum bmep 10.5 barMoments of inertia Ie = 0.25 kgm2

. Id = 0.30 kgm2

Table 9.4 indicates a service factor of 4.8, giving a design torque of110× 4.8= 530Nm.

186 Engine Testing

Table 9.4 Service factors for dynamometer/engine combinations

Dynamometer

type

Number of cylinders

Diesel Gasolene

1/2 3/4/5 6 8 10+ 1/2 3/4/5 6 8 10+

Hydraulic 4.5 4.0 3.7 3.3 3.0 3.7 3.3 3.0 2.7 2.4Hyd.+ dyno.start

6.0 5.0 4.3 3.7 3.0 5.2 4.3 3.6 3.1 2.4

Eddy current(EC)

5.0 4.5 4.0 3.5 3.0 4.2 3.8 3.3 2.9 2.4

EC+ dyno.start

6.5 5.5 4.5 4.0 3.0 5.7 4.8 3.8 3.4 2.4

d.c.+ dyno.start

8.0 6.5 5.0 4.0 3.0 7.2 5.8 4.3 3.4 2.4

It is proposed to connect the two couplings by a steel shaft of the followingdimensions:

Diameter D= 40mmLength L= 500mmModulus of rigidity G= 80× 109 Pa

From eq. (9a), torsional stress = 42Mpa, very conservative.

From eq. (10a)

Cs =�×0�044×80×109

32×0�5= 40200Nm/rad

Consider first the case when rigid couplings are employed:

nc =60

2�

√

40200×0�55

0�25×0�30= 5185c�p�m�

For a four cylinder, four-stroke engine, we have seen that the first major criticaloccurs at order N0 = 2, corresponding to an engine speed of 2592 rev/min. This fallsright in the middle of the engine speed range and is clearly unacceptable. This is atypical result to be expected if an attempt is made to dispense with flexible couplings.

The resonant speed needs to be reduced and it is a common practice to arrangefor this to lie between either the cranking and idling speeds or between the idlingand minimum full load speeds. In the present case these latter speeds are 500 and


1000 rev/min, respectively. This suggests a critical speed Nc of 750 rev/min and acorresponding resonant frequency nc = 1500 cycles/min.

This calls for a reduction in the torsional stiffness in the ratio:(

1500

5185

)2

i.e. to 3364Nm/rad.The combined torsional stiffness of several elements in series is given by:

1

C=

1

C1

+1

C2

+1

C3

+· · · (14)

This equation indicates that the desired stiffness could be achieved by the use of twoflexible couplings each of stiffness 7480Nm/rad. A manufacturer’s catalogue showsa multi-bush coupling having the following characteristics:

Maximum torque 814Nm (adequate)Rated torque 170NmMaximum continuous vibratory torque ±136NmShore (IHRD) hardness 50/55Dynamic torsional stiffness 8400Nm/rad

Substituting this value in eq. (14) indicates a combined stiffness of 3800Nm/rad. Sub-stituting in eq. (1) gives nc = 1573, corresponding to an engine speed of 786 rev/min,which is acceptable.

It remains to check on the probable amplitude of any torsional oscillation at thecritical speed. Under no-load conditions, the imep of the engine is likely to be in theregion of 2 bar, indicating, from eq. (5a), a mean turning moment Mmean = 8Nm.

From Table 9.1, p factor= 1.91, giving Tex = 15Nm per cylinder:

∑

Tex = 4×15= 60Nm

Table 9.3 indicates a dynamic magnifier M= 10.5, the combined dynamic mag-nifier from eq. (13)= 7.4.

The corresponding value of the vibratory torque, from eq. (7), is then:

Tv =60×7�4

�1+0�25/0�30�=±242 Nm

This is in fact outside the coupling continuous rating of ±136Nm, but multiple bushcouplings are tolerant of brief periods of quite severe overload and this solutionshould be acceptable provided the engine is run fairly quickly through the criticalspeed. An alternative would be to choose a coupling of similar stiffness using SBRbushes of 60/65 hardness. Table 9.3 shows that the dynamic magnifier is reducedfrom 10.5 to 2.7, with a corresponding reduction in Tv.

188 Engine Testing

If in place of an eddy current dynamometer we were to employ a d.c. machine,the inertia Ib would be of the order of 1 kgm2, four times greater.

This has two adverse effects:

1. Service factor, from Table 9.4 increased from 4.8 to 5.8;2. The denominator in eq. (7) is reduced from (1+ 0.25/0.30)= 1.83 to

(1+ 0.25/1.0)= 1.25, corresponding to an increase in the vibratory torque for agiven exciting torque of nearly 50 per cent.

This is a general rule: the greater the inertia of the dynamometer the more severe

the torsional stresses generated by a given exciting torque.

An application of eq. (1) shows that for the same critical frequency the combinedstiffness must be increased from 3364Nm/rad to 5400Nm/rad. We can meet thisrequirement by changing the bushes from Shore Hardness 50/55 to Shore Hardness60/65, increasing the dynamic torsional stiffness of each coupling from 8400Nm/radto 14000Nm/rad (in general, the usual range of hardness numbers, from 50/55 to75/80, corresponds to a stiffness range of about 3:1, a useful degree of flexibility forthe designer).

Eq. (1) shows that with this revised coupling stiffness nc changes from 1573cycles/min to 1614 cycles/min, and this should be acceptable. The oscillatory torquegenerated at the critical speed is increased by the two factors mentioned above,but reduced to some extent by the lower dynamic magnifier for the harder rubber,M= 8.6 against M= 10.5. As before, prolonged running at the critical speed shouldbe avoided.

For completeness, we should check the whirling speed from eq. (11). The massof the shaft per unit length is: Ws = 9.80 kg/m.

Nw =30�

0�502

√

200×109×�×0�044

64×9�80= 19100 r�p�m�

The mass of the shaft+ half couplings is found to be 12 kg and the combinedradial stiffness 33.6MN/m. From eq. (12a):

Nt =30

�

√

33�6×106

12= 16000 r�p�m�

then from eq. (12b), whirling speed= 12 300 rev/min, which is satisfactory.Note, however, that, if shaft length were increased from 500 to 750mm, whirling

speed would be reduced to about 7300 rev/min, which is barely acceptable. This is acommon problem, usually dealt with by the use of tubular shafts, which have muchgreater transverse stiffness for a given mass.

There is no safe alternative, when confronted with an engine of which the charac-

teristics differ significantly from any run previously on a given test bed, to following

through this design procedure.


Figure 9.10 Annular type rubber coupling

An alternative solution

The above worked example makes use of two multiple-bush type rubber couplingswith a solid intermediate shaft. An alternative is to make use of a conventionalpropeller shaft with two universal joints, as used in road vehicles, with the additionof a coupling incorporating an annular rubber element in shear to give the necessarytorsional flexibility. These couplings, Fig. 9.10, are generally softer than the multiplebush type for a given torque capacity, but are less tolerant of operation near a criticalspeed. If it is decided to use a conventional universal joint shaft, the supplier shouldbe informed of the maximum speed at which it is to run. This will probably bemuch higher than is usual in the vehicle and may call for tighter than usual limits onbalance of the shaft.

Shock loading of couplings due to cranking, irregular running and torquereversal

Systems for starting and cranking engines are described in Chapter 8, where it isemphasized that during engine starting severe transient torques can arise. These have

190 Engine Testing

been known to result in the failure of flexible couplings of apparently adequatetorque capacity. The maximum torque that can be necessary to get a green engineover t.d.c. or that can be generated at first firing should be estimated and checkedagainst maximum coupling capacity.

Irregular running or imbalance between the powers generated by individual cylin-ders can give rise to exciting torque harmonics of lower order than expected in amulticylinder engine and should be borne in mind as a possible source of rough run-ning. Finally, there is the possibility of momentary torque reversal when the enginecomes to rest on shutdown.

However, the most serious problems associated with the starting process arisewhen the engine first fires. Particularly when, as is common practice, the engine ismotored to prime the fuel injection pump, the first firing impulses can give rise tosevere shocks. Annular type rubber couplings, Fig. 9.10, can fail by shearing underthese conditions. In some cases, it is necessary to fit a torque limiter or slippingclutch to deal with this problem.

Axial shock loading

Engine test systems that incorporate automatic shaft docking systems have to providefor the axial loads on both the engine and dynamometer imposed by such a system.In some high volume production facilities, an intermediate pedestal bearing isolatesthe dynamometer from both the axial loads of normal docking operation and forcases when the docking splines jam ‘nose to nose’; in these cases the docking controlsystem should be programmed to back off the engine, spin the dynamometer andretry the docking.

Selection of coupling torque capacity

Initial selection is based on the maximum rated torque with consideration given tothe type of engine and number of cylinders, dynamometer characteristics and inertia.Table 9.4, reproduced by courtesy of Twiflex Ltd, shows recommended service fac-tors for a range of engine and dynamometer combinations. The rated torque multipliedby the service factor must not exceed the permitted maximum torque of the coupling.

Other manufacturers may adopt different rating conventions, but Table 9.4 givesvaluable guidance as to the degree of severity of operation for different situations.Thus, for example, a single cylinder diesel engine coupled to a d.c. machine withdynamometer start calls for a margin of capacity three times as great as an eightcylinder gasoline engine coupled to a hydraulic dynamometer.

Figure 9.11 shows the approximate range of torsional stiffness associated withthree types of flexible coupling: the annular type as illustrated in Fig. 9.10, themultiple bush design of Fig. 9.8 and a development of the multiple bush design whichpermits a greater degree of misalignment and makes use of double-ended bushedlinks between the two halves of the coupling. The stiffness figures of Fig. 9.11 referto a single coupling.


Torque capacity (Nm)

Twin rigid universaljoints plus annulartype coupling as Fig. 9.10

Rubber bush coupling,twin brushes joined byswinging links

Rubber bush coupling

Dynam

ic tors

ional stiffness (N

m/r

ad)

109

108

107

106

105

104

103

102

106 107104 105103102

Figure 9.11 Ranges of torsional stiffness for different types of rubber coupling

The role of the engine clutch

Vehicle engines are invariably fitted with a clutch and this may or may not be retainedon the test bed. The advantage of retaining the clutch is that it acts as a torquelimiter under shock or torsional vibration conditions. The disadvantages are that itmay creep, particularly when torsional vibration is present, leading to ambiguitiesin power measurement, while it is usually necessary, when the clutch is retained, toprovide an outboard bearing. Clutch disc springs may have limited life under testbed conditions.

Balancing of drive line components

This is a matter which is often not taken sufficiently seriously and can lead toa range of troubles, including damage (which can be very puzzling) to bearings,unsatisfactory performance of such items as torque transducers, transmitted vibrationto unexpected locations and serious drive line failures. Particular care should be takenin the choice of couplings for torque shaft dynamometers: couplings such as themultiple disc type, Fig. 9.7, cannot be relied upon to centre these devices sufficientlyaccurately.

192 Engine Testing

It has sometimes been found necessary to carry out in situ balancing of a compositeengine drive line where the sum of the out of balance forces in a particular radialrelationship causes unacceptable vibration; specialist companies with mobile plantexist to provide this service.

Conventional universal joint type cardan shafts are often required to run at higherspeeds in test bed applications than is usual in vehicles; when ordering, the maximumspeed should be specified and, possibly, a more precise level of balancing than thestandard specified.

BS 5265, Mechanical balancing of rotating bodies,9 gives a valuable discussion ofthe subject and specifies 11 Balance Quality Grades. Drive line components shouldgenerally be balanced to Grade G 6.3, or, for high speeds, to grade G 2.5. Thestandard gives a logarithmic plot of the permissible displacement of the centre ofgravity of the rotating assembly from the geometrical axis against speed. To give anidea of the magnitudes involved, G 6.3 corresponds to a displacement of 0.06mm at1000 rev/min, falling to 0.01mm at 5000 rev/min.

Alignment of engine and dynamometer

This is a fairly complex and quite important matter. For a full treatment anddescription of alignment techniques, see Ref. 4. Differential thermal growth and themovement of the engine on its flexible mountings when on load should be takeninto account and if possible the mean position should be determined. The laser-based alignment systems now available greatly reduce the effort and skill requiredto achieve satisfactory levels of accuracy. In particular, they are able to bridgelarge gaps between flanges without any compensation for droop and deflection ofarms carrying dial indicators, a considerable problem with conventional alignmentmethods.

There are essentially three types of shaft having different alignment requirementsto be considered:

1. Rubber bush and flexible disc couplings should be aligned as accurately aspossible as any misalignment encourages heating of the elements and fatiguedamage.

2. Gear type couplings require a minimummisalignment of about 0.04� to encouragethe maintenance of an adequate lubricant film between the teeth.

3. Most manufacturers of universal joint propeller shafts recommend a small degreeof misalignment to prevent brinelling of the universal joint needle rollers. Notethat it is essential, in order to avoid induced torsional oscillations, that the twoyokes of the intermediate shaft joints should lie in the same plane.

Distance between end flanges can be critical, as incorrect positioning can lead tothe imposition of axial loads on bearings of engine or dynamometer.


Guarding of coupling shafts

Not only is the failure of a high speed coupling shaft potentially dangerous, as verylarge amounts of energy can be released, it is quite common. The ideal shaft-guardwill contain the debris resulting from a failure of any part of the drive line andprevent accidental or casual contact with rotating parts, while being sufficiently lightand adaptable not to interfere with engine rigging and alignment. A guard systemthat is very inconvenient to use will eventually be misused or neglected.

A really substantial guard, preferably a steel tube not less than 6mm thick, splitand hinged in the horizontal plane for shaft access, is an essential precaution. Thehinged ‘lid’ should be interlocked with the emergency stop circuit to prevent enginecranking or running while it is open. Many designs include shaft restraint devicesloosely fitting around the tubular portion, made of wood or a non-metallic compositeand intended to prevent a failing shaft from whirling; these should not be so closeto the shaft as to be touched by it during normal starting or running otherwise theywill be the cause of failure rather than a prevention of damage.

Engine to dynamometer coupling: summary of designprocedure

1. Establish speed range and torque characteristic of engine to be tested. Is itproposed to run the engine on load throughout this range?

2. Make a preliminary selection of a suitable drive shaft. Check that maximumpermitted speed is adequate. Check drive shaft stresses and specify material.Look into possible stress raisers.

3. Check manufacturer’s recommendations regarding load factor and otherlimitations.

4. Establish rotational inertias of engine and dynamometer and stiffness of proposedshaft and coupling assembly. Make a preliminary calculation of torsional criticalspeed from eq. (1). (In the case of large multicylinder engines consider makinga complete analysis of torsional behaviour.)

5. Modify specification of shaft components as necessary to position torsionalcritical speeds suitably. If necessary, consider use of viscous torsional dampers.

6. Calculate vibratory torques at critical speeds and check against capacity of shaftcomponents. If necessary specify ‘no go’ areas for speed and load.

7. Check whirling speeds.8. Specify alignment requirements.9. Design shaft guards.

Flywheels

No treatment of the engine/dynamometer drive line would be complete withoutmention of flywheels which may form a discrete part of the shaft system. A flywheel

194 Engine Testing

is a device that stores kinetic energy. The energy stored may be precisely calculatedand is instantly available. The storage capacity is directly related to the mass momentof inertia which is calculated by:

I = k×M×R2

where:

I=moment of inertia (kgm2)k= inertia constant (dependent upon shape)m=mass (kg)R= radius of flywheel mass

In the case of a flywheel taking the form of a uniform disk, which is the commonform found within dynamometer cells and chassis dynamometer designs:

I =1

2MR2

The engine or vehicle test engineer would normally expect to deal with flywheels intwo roles:

1. As part of the test object, as in the common case of an engine flywheel whereit forms part of the engine/dynamometer shaft system and contributes to thesystem’s inertial masses taken into account during a torsional analysis.

2. As part of the test equipment where one or more flywheels may be used toprovide actual inertia that would, in ‘real life’, be that of the vehicle or somepart of the engine driven system.

No mention of flywheels should be made without consideration of the safety ofthe application. The uncontrolled discharge of energy from any storage device ishazardous. The classic case of a flywheel failing by bursting is now exceptionally rareand invariably due to incompetent modification rather than the nineteenth centuryproblems of poor materials, poor design or overspeeding.

In the case of engine flywheels, the potential danger in the test cell is the shaftsystem attached to it. This may be quite different in mass and fixing detail from itsfinal application connection, and can cause overload leading to failure. Cases are onrecord where shock loading caused by connecting shafts touching the guard systemdue to excessive engine movement has created shock loads that have led to the castengine flywheel fracturing, with severe consequential damage.

The most common hazard of test rig mounted flywheels is caused by bearing orclutch failure where consequential damage is exacerbated by the considerable energyavailable to fracture connected devices or because of the time that the flywheel andconnected devices will rotate before the stored energy is dissipated and movementis stopped.

It is vital that flywheels are guarded in such a manner as to prevent absolutelyaccidental entrainment of clothing or cables, etc.


A common and easy to comprehend use of flywheels is as part of a vehicle braketesting rig. In these devices, flywheels supply the energy that has to be absorbed anddissipated by the brake system under test. The rig motor is only used to acceleratethe chosen flywheel combinations up to the rotational speed required to simulate thevehicle axle speed at the chosen vehicle speed. Flywheel brake rigs have been madeup to the size that can provide the same kinetic energy as fully loaded high speedtrains. Flywheels are also used on rigs used to test automatic automotive gearboxes.

Test rig flywheel sets need to be rigidly and securely mounted and balanced to thehighest practical standard. Multiples of flywheels forming a common system that canbe engaged in different combinations and in any radial relationship require particularcare in the design of both their base frame and individual bearing supports. Suchsystems can produce virtually infinite combinations of shaft balance and require eachindividual mass to be as well balanced and aligned on as rigid a base as possible.

Simulation of inertia∗ versus iron inertia

Modern a.c. dynamometer systems and control software have significantly replacedthe use of flywheels in chassis and engine dynamometer systems in the automotiveindustry. Any perceived shortcoming in the speed of response or accuracy of thesimulation is usually considered to be of less concern than the mechanical simplicityof the electric dynamometer system and the reduction in required cell space.

Finally, it should be remembered that, unless engine rig flywheels are able to beengaged through a clutch then the engine starting/cranking system will have to becapable of accelerating engine, dynamometer and flywheel mass up to engine startspeed.

Notation

Frequency of torsional oscillation n cycles/minCritical frequency of torsional oscillation nc cycles/minStiffness of coupling shaft Cs Nm/radRotational inertia of engine Ie kgm

2

Rotational inertia of dynamometer Ib kgm2

Amplitude of exciting torque Tx NmAmplitude of torsional oscillation � radStatic deflection of shaft �0 radDynamic magnifier M

Dynamic magnifier at critical frequency Mc

Order of harmonic component No

∗ Some readers may object to the phrase ‘simulation of inertia’ since one is simulating theeffects rather than the attribute, but the concept has wide industrial acceptance.

196 Engine Testing

Number of cylinders Ncyl

Mean turning moment Mmean NmIndicated mean effective pressure pi barCylinder bore B mmStroke S mmComponent of tangential effort Tm NmAmplitude of vibratory torque Tv NmEngine speed corresponding to nc Nc rev/minMaximum shear stress in shaft N/m2

Whirling speed of shaft NW rev/minTransverse critical frequency Nt cycles/minDynamic torsional stiffness of coupling Cc Nm/radDamping energy ratio �

Modulus of elasticity of shaft material E PaModulus of rigidity of shaft material G Pa(for steel, E= 200× 109 Pa, G= 80× 109 Pa)

References

1. Den Hartog, J.P. (1956) Mechanical Vibrations, McGraw-Hill, Maidenhead, UK.2. Ker Wilson, W. (1963) Practical Solution to Torsional Vibration Problems (5 vols),

Chapman and Hall, London.3. Ker Wilson, W. (1959) Vibration Engineering, Griffin, London.4. Neale, M.J., Needham, P. and Horrell, R. (1998) Couplings and Shaft Alignment, Mecha-

nical Engineering Publications, London.5. Rulebook, Chapter 8. Shaft vibration and alignment, Lloyd’s Register of Shipping, London.6. Pilkey, W.D. (1997) Peterson’s Stress Concentration Factors, Wiley, New York.7. Young, W.C. (1989) Roark’s Formulas for Stress and Strain, McGraw-Hill, New York.8. BS 6613 Methods for Specifying Characteristics of Resilient Shaft Couplings.9. BS 5265 Parts 2 and 3 Mechanical Balancing of Rotating Bodies.

Further reading

BS 4675 Parts 1 and 2 Mechanical Vibration in Rotating Machinery.BS 6861 Part 1 Method for Determination of Permissible Residual Unbalance.BS 6716 Guide to Properties and Types of Rubber.Nestorides, E.J. (1958) A Handbook of Torsional Vibration, Cambridge University

Press, Cambridge.

10 Electrical design considerations

Introduction

The electrical system of a test facility provides the power, nerves and operating logicthat control the test piece, the test instrumentation and modules of building services.The power distribution to, and integration of, these many parts falls significantlywithin the remit of the electrical engineer, whose drawings will be used as theprimary documentation in system commissioning or any subsequent fault findingtasks. Integration of test cell modules equipment built to American standards inEurope and the rest of the world and vice versa can create detailed interface problemsnot appreciated by specialists at either end of the project unless they have experiencein dealing with each other’s practices.

The theme of system integration is nowhere more pertinent than within the roleof the electrical designer.

Note: In this text and in context with electrical engineering, the UK English terms

‘earth’ and ‘earthing’ have been generally used in preference to the US English

terms ‘ground’ and ‘grounding’ with which they are, in the context of this book,

interchangeable.

The electrical engineer’s design role

The electrical engineer’s prime guidance are the operational and functional speci-fications (see Chapter 1) from which he will develop the final detailed functionalspecification, system schematics, power distribution and alarm logic matrices.

In the design process much of the mechanical and civil plant has to be identifiedbefore the electrical engineer can start to calculate the electrical power requiredand produce an electrical distribution scheme, complete with the control cabinetsand interconnecting cable ways; all of which may have implications for the otherspecialist designers. System integration is therefore an iterative process with theelectrical engineer as a key team member.

Engine test facilities are, by the nature of their component parts and their role,particularly vulnerable to signal distortion due to various forms of electrical noise.To avoid signal corruption and cell downtime due to instrumentation errors, specialand detailed attention must be made to the standard of electrical installation within

198 Engine Testing

a test facility. The electrical designer and installation supervisor must be able to takea holistic approach and be aware of the need to make an electrically integrated andmutually compatible system. Facilities that are developed on an ad hoc basis oftenfall foul of unforeseen signal interference or control logic errors within interactingsubsystems.

General characteristics of the electrical installation

Perhaps more than any other aspect of test cell design and construction the electricalinstallation is subject to regulations, most of which have statutory force. It is essentialthat any engineer responsible for the design or construction of a test cell in the UKshould be familiar with BS 7671 Requirements for Electrical Installations. There arealso many British Standards specifying individual features of an electrical system;other countries will have their own national standards. In the USA the relevantelectrical standards may be accessed through the ANSI website.

The rate of change of regulations, particularly European regulation, will outpacethat of any general text book (Table 10.1). (BS EN) 60204 is of particular relevanceto test cell design since it includes general rules concerning safety interlocking andthe shutting down of rotating plant as in section 9.2.2 covering shutting down ofengine/dynamometer combinations (see Chapter 11, section Safety systems).

While these regulations cover most aspects of the electrical installation in test cells,there are several particular features that are a consequence of the special conditionsassociated with the test cell environment which are not explicitly covered. These willbe included within this chapter, as will the special electrical design features requiredif four quadrant electrical dynamometers (see Chapter 8) are being included withinthe facility.

Table 10.1 Electrical installation regulations

IEC 60204–1/97 Safety of machinery – electrical equipment ofmachinesEN 60204–1/97

IEC 1010–1 Safety requirement for electrical equipment formeasurement, control and laboratory useEN 61010–1/A2

EN 50178/98 Electronic equipment of power installationIEC 61800–3/96 Adjustable speed electrical power drive systemsEN 61800–3/96 EMC product standardEN 61000–6–4 Electromagnetic influence – emissionEN 61000–6–2 Electromagnetic influence – immunityEN 61010–1 Safety requirements of laboratory (test cell)

equipmentIEC 61010–1 designed for measurement and control

Electrical design considerations 199

Physical environment

The physical environment inside a test cell can vary between the extremes used totest the engine or vehicle and the test cell designer must ensure that the test plant israted such that it can withstand the full range of operating conditions specified.

The control room and services spaces will, at worst, experience the range ofambient conditions met at the geographical location, which can exceed the ‘normal’conditions expected by some instrument manufacturers. If air conditioning of controlroom or services spaces is installed in atmospheres of high humidity, the electricaldesigner must consider protection against condensation within control cabinets anddynamometer drive cabinets.

The guidelines for acceptable ambient conditions for most electrical plant are asfollows:

• The ambient temperature of the control cabinet should be between +5�C and+35�C.

• The ambient temperature for a control room printer is restricted to +35�C andfor the PC to +38�C.

• The ambient temperature of the test cell should be between +5�C and +45�C.• The air must be free of abnormal amounts of dust, acids, corroding gases and

salt (use filter).• The relative humidity at +40�C must not be more than 50 and 90 per cent at

+20�C (use heating).

The rating of electrical devices, such as motors and enclosures, must be correctlydefined at the design stage.

In some vehicle and engine cells, particularly large diesel and engine rebuildfacilities, it is common practice for a high pressure heated water washer to be usedin which case the test bed enclosures should have a rating of at least IP×5 (seeTable 10.2 below).

Electrical signal and measurement interference

The protection methodology against signal interference in engine test cells haschanged significantly over the period since the mid-1990s because of the arrival ofelectromagnetic interference in the radio frequencies as the dominating source ofsignal corruption.

The pulse width modulating drive technology based on fast switching insulated

bipolar gate transistors (IGBT) devices used with a.c. dynamometers has reduced thetotal harmonic distortion (THD) experienced in power supplies from that approach-ing the 30 per cent that was produced by a d.c. thyristor controlled drive to <5per cent THD.

200 Engine Testing

Table 10.2 Ingress protection rating details

1stDigit

2ndDigit

Protection from moistureProtection from solid objects

IP54 = IP Letter Code IP5

41st Digit2nd Digit

Non-protected

Protected againstsolid objects greater

than 50 mm

Protected againstdripping water

Protected againstdripping water when

tilted up to 15°

Protected againstspraying water

Protected againstsplashing water

Protected againstwater jets

Protected againstheavy seas

Protected againstthe effects of

immersion

Protected againstsubmersion


than 12 mm


than 2.5 mm∅


than 1.0 mm∅

Dust protected

Dust tight

0

1

2

3

4

5

6

1

2

3

15°

60°

4

5

6

7

8

0 Non-protected


I

130 A

Harmonic

25 A

1 2 3 4 5 6 7 8 9 ...

3. Basic oscillation

Signal accordingspectrum analyser

Frequency50 Hz 150 Hz

Order

10 A

Sinusoidal basic oscillation (e.g. 50 Hz)

Oscilloscope display

t

I

3. Harmonic (sinus oscillation,

f = 3 × 50 Hz = 150 Hz)

Figure 10.1 Harmonic distortion

THD%=

√

√

√

k∑

2

(

Hx

Hf

)

2

where Hx is the amplitude of any harmonic order and Hf is the amplitude of thefundamental harmonic (1st order). Most sensitive devices should be unaffected by aTHD of less than 8 per cent.

The harmonic distortion produced by thyristor drives associated d.c. machines isload-dependent and causes a voltage drop at the power supply (see Fig. 10.1).

While IGBT technology has reduced THD that caused problems in facilities fittedwith d.c. dynamometers, it has joined other digital devices in introducing disturbancein the frequency range of 150 to 30MHz.

IBGT drive systems produce a common mode, load-independent disturbance thatcauses unpredictable flow through the facility earth system.

This has meant that standard practices concerning signal cable protection andprovision of ‘clean earth’ connections that are separate from ‘protective earths’ aretending to change in order to fight the new enemy: electromagnetic interference(EMI). The connection of shielded cables between devices shown in Fig. 10.2 was the

Device 1 Device 2

Pig tail connectionat one end of shield

Ground loop current possible ifshield is connected at both ends

∆V

Figure 10.2 Recommended method of connecting shielded cable betweendevices to prevent ground loop current distortion of signals

202 Engine Testing

Shielded braid connected roundcomplete circumference at

both ends

Compensating cable

Low impedance earth

EMI

Device 2Device 1

Figure 10.3 Recommended method of connecting shielded cable betweendevices to prevent EMI and ground loop corruption of signals

standard method aimed at preventing earth loop distortion of signals in the shieldedcable. Earth loops are caused by different earth potential values occurring across ameasuring or signal circuit which induces compensating currents.

This single end connection of the cable shield shown in Fig. 10.2 offers noprotection from EMI which requires connection at both ends (at both devices). Thisrequires that the ground loops are defeated by different measures. At the level ofshielded cable connection the countermeasure is to run a compensation lead of highsurface area in parallel with the shield as shown in Fig. 10.3. This method alsoreduces the vulnerability of the connection to external magnetic flux fields.

Earthing system design

The earthing systems used in many industrial installations are primarily designedas a human protection measure. However, electromagnetic compatibility (EMC)requirements increasingly require high frequency equipotential bonding, achieved bycontinuous linking of all ground potentials. This modern practice of earthing deviceshas significantly changed layouts in which there was once a single protective earth(PE) connection, plus when deemed necessary, a ‘clean earth’ physically distant forthe PE. A key feature of modern EMC practice is the provision of multiple earthconnections to a common earthed grid of the lowest possible impedance. The physicaldetails of such connection practice are shown in Figs 10.4 and 10.5.

To ensure the most satisfactory functioning of these EMC systems, the earthingsystem needs to be incorporated into a new building design and the specification ofthe electrical installation. Ideally the building should be constructed with a groundmat made of welded steel embedded in the concrete floor and a circumferential earthstrap.


Dyno.Equipment

Multipoint earthing –Welded steel mats

embedded in concrete

Galvanized ringearth embedded infoundation concrete

>30 × 3 mm

Equipotentialearth strip

Equipotentialearth strip

Figure 10.4 EMC equipotential earthing grid incorporated in new building

The layout of cabling

Transducer signals are usually ‘conditioned’ as near to the transducer as possible;nevertheless, the resultant conditioned signals are commonly in the range 0 to 10Vd.c. or 0 to 5mA, very small when compared with the voltage differences and currentflows that may be present in power cables in the immediate vicinity of the signallines. Cable separation and layout is of vital importance. Practices that create chaoticjumbles of cables beneath trench covers and ‘spare’ cable lengths looped or coiledin the base of distribution boxes are simply not acceptable in modern laboratories.

Most engine test facilities will contain the following types of wiring which havedistinctly different roles and which must be prevented from interfering with eachother:

• power cables, mains supply ranging from high power wiring for dynamometersthrough three-phase and single-phase distribution for services and instrumentsto low power supplies for special transducers, also high current d.c. supply forstarter systems;

• control cables for inductive loads, relays, etc.;• signal cables:

– digital control with resistive load;– ethernet, RS232, RS422, IEEE1394;

204 Engine Testing

PE wire

Grid

Connection to pipe

Welded polt

Electricalconducting

strap

Figure 10.5 Continuous linked network of all ground potentials

– bus systems, such as CAN;– 24 Vd.c. supplies;

• measuring cables associated with transducers and instrumentation transmittinganalogue signals.

In the following paragraphs, the common causes of signal interference are identifiedand practices recommended to avoid the problems described.

Inductive interference

Inductive interference is caused by the magnetic flux generated by electrical currentsinducing voltages in nearby conductors.

Countermeasures include:

• Do not run power cables close to control or signal cables (also see Capacitiveinterference).


Power cables Control cables

Instrumentation cables

Signal cables screened

Signal cables unscreened

Galvanized steel trayor segmented trunking

Figure 10.6 Segmented trays or trunking and separation of types of cable

• Use either segmented trunking or different cable tray sections as shown inFig. 10.3.

• Use twisted pair cables for connection of devices requiring supply and returnconnection (30 twists per metre will reduce interference voltage by a factor of 25).

• Use shielded signal cables and connect the shield to earth at both ends.

Figure 10.6 below shows a suggested layout of cables running in open cable trays.The same relationship and metal division between cable types may be achievedusing segmented trunking, trays or ladders. It is important that the segments of suchmetallic support systems are connected together as part of the earth bonding system;it is not acceptable to rely on metal to metal contact of the segments.

Note that cables of different types crossing through metal segments at 90� to themain run do not normally cause problems.

Capacitive interference

Capacitive interference can occur if signal cables with different voltage levels arerun closely together. It can also be caused by power cables running close to signallines.


• separate signal cables with differing voltage levels;• do not run signal cables close to power cables;• if possible use shielded cable for vulnerable signal lines.

206 Engine Testing

Electromagnetic interference

Electromagnetic interference can induce both currents and voltages in signal cables.It may be caused by a number of ‘noise’ transmitters ranging from spark plugs tomobile communication devices.

RF noise produced by motor drives based on IBGT devices is inherent in thetechnology with inverter frequencies of between 3 and 4 kHz and is due to the steepleading and failing edge of the, typically 500V, pulses that have a voltage rise ofaround 2 kV/�sec. There may be some scope in varying the pattern of frequencydisturbances produced at a particular drive/site combination by adjustment of thepulsing frequency; the supplier would need to be consulted.

Countermeasures for a.c. drive-induced noise include:

• Choice and layout of both motor/drive cabinet and drive cabinet to supply cablesis very important:

– Motor cable should have the three multicore power conductors in a symmet-rical arrangement within common braided and foil screens. The bonding ofthe screen at the termination points must contact 360� of the braid and be oflow impedance.

– The power cables should contain symmetrically arranged conductors of lowinductivity within a concentric PE conductor.

– Single core per phase cables should be laid as shown in Fig. 10.7.

• Signal cables should be screened.• Signal cables should be encased in metal trunking or laid within metal cable tray.

D

L2

L3

PE

L1

L3

L2

PE

L1

L3

PE

L2

L1

D D

D 2 × D

Multicore cable layout on trench ortray

Double layeringSingle layering

Figure 10.7 Recommended spacing and layout of power cables


The layout of power cables of both multicore and single core is shown in Fig. 10.7following recommended practice.

Conductive coupling interference

Conductive coupling interferencemay occur when there is a supply voltage differencebetween a number of control or measuring devices. It is usually caused by longsupply line length or inappropriate distribution layout.


• Keep supply cables short and of sufficient conductor size to minimize voltage drop.• Avoid common return lines for different control or measurement devices by

running separate supply lines for each device.

The same problem may occur when two devices are fed from different power supplies.To minimize interference in analogue signal cables, an equipotential bonding cable

or strap should connect the two devices running as close as possible to the signalcable. It is recommended that the bonding cable resistance be <1/10 (one tenth) ofthe cable screen resistance.

Integration of a.c. dynamometer systems

Careful consideration should be made when integrating an a.c drive system intoan engine test facility. Where possible it is most advisable to provide a dedicatedelectrical supply to an a.c drive or number of a.c. drives.

There needs to be a clear understanding of the status of the existing electri-cal supply network, and the work involved in providing a new supply for an a.cdynamometer system. It is important to calculate the correct rating for a new supplytransformer. This is a highly specialized subject and the details may change depend-ing on the design of the supply system and a.c. devices, but the general rule is thatthe mains short circuit apparent power (SSC) needs to be at least 20 times greaterthan the nominal apparent power of one dynamometer (SN). Hence,

SSC/SN > 20�

In a sample calculation based on a single 220 kW dynamometer, the following sizingof the supply transformer may be found:

PN = 220 kW nominal power, cos �= 1, overload 25%SN = PN× cos �= 220 kVASSC = SN×20= 4400 kVAStandard transformer (ST) ∼Usc = 6%

208 Engine Testing

ST = SSC×0�06= 264 kVA ⇒ Transformer= 315 kVAST = SN×1�25= 275 kVA ⇒ Transformer= 315 kVA.

In the case of multiple dynamometer installation, some manufacturers in some con-ditions will allow a reduction in the SSC/SN ratio; this can reduce the cost of theprimary supply.

Shown below is the basic calculation of power supply transformer in the case of3× 220 kW dynamometers:

PN = 220 kW nominal power, cos �= 1, overload 25 per cent, diversity 0.7SN = 3 × PN× cos �= 660 kVASSC1 = SN×20= 13 200 kVA in the case of no reduction for multiple machinesSSC2 = SN×10= 6600 kVA in the case of ratio reduction to minimumStandard transformer ∼Usc = 6%ST = SSC1×0�06= 792 kVA ⇒ Transformer = 800 kVAST = SSC2×0�06= 396 kVA ⇒ Transformer = 400 kVAST = SN×1�25×0�7= 577 kVA ⇒ Transformer = 630 kVA.

The ratio reduction and diversity (0.7) will reduce the transformer size from 800 kVAto 630 kVA.

Supply interconnection of disturbing and sensitive devices

An infinite variation of transformer and connection systems is possible, but Fig. 10.8shows the range from poor (a) to a recommended layout (c).

Common bus

Commonearth!

Sensitivedevices

(a) (b) (c)

Disturbingdevice

Figure 10.8 Different power connection layouts ranging from poor (a) torecommended (c)


The worst connection scheme is shown in Fig. 10.8a, where both the sensitive anda disturbing device, such as a PWM drive, are closely connected on a common localbus. Figure 10.8b shows an improved version of (a) where the common bus is localto the transformer rather than the devices. The best layout is based on Fig. 10.8c,where the devices are fed from separate transformers that share a common earthconnection.

Power utility representatives drawn in projects involving a.c. dynamometer orchassis dynamometer installations may not at first appreciate that, for the majorityof their working life, these machines are not operating as motors in the usual sensebut are working as engine/vehicle power absorbers, or in electrical terms they areexporting, rather than importing, electrical power. It may be necessary to use theexpertise of the equipment suppliers at the planning stage if upgraded power systemsare being provided.

Electrical power supply specification

Most suppliers of major plant will include the power supply conditions their plantwill accept; in the case of a.c. dynamometers this may require a dedicated isolatingtransformer to provide both isolation and the required supply voltage.

The EEC Directive (89/336/EEC) gives the acceptable limits of high frequencydistortion on mains supply; the categories EN 55011 QP A2∗ and A2 are applicableto supplies for a.c. automotive dynamometers.

A typical mains power specification in the UK is shown in Table 10.3.

Fire stopping of cable penetrations in cell structure

Where cables or cable trunking break through the test cell walls, roof or floor theymust pass through a physical ‘fire block’. There are some designs of cable wall-box that allow the fire block to be disassembled and extra cables added; none ofthese are particularly easy to use after a year or more in service and, if used, itis advisable to build in a number of spare cables. Since most cells have a firerating of at least one hour it is important not to compromise this by casual misuse,poor maintenance or inappropriate cable ways through the structure. A hole in thecontrol room wall loosely stuffed with bags of intumescent material (fire barrier)is not adequate protection for cells fitted with gas-based fire suppression systems,particularly CO2, because the pressure wave created by gas release will be capable ofblowing out such stuffing unless it is correctly retained by a wall plate on either side.

Plastic ‘soil pipes’ cast into the floor are a convenient way of carrying signalcables between test cell and control room. These ducts should be laid to a fallin the direction of the cell and should have a raised lip to prevent drainage ofliquid into them. Spare cables should be laid during installation. These pipes can be

210 Engine Testing

Table 10.3 Typical UK electrical supply specification

Voltage 230/240 VAC ± 10%

Frequency 0.99 to 1.01 of nominal frequency 50 or 60Hzcontinuously

0.98 to 1.02 of nominal frequency 50 or 60Hz shorttime

Harmonics distortion Harmonics distortion is not to exceed 10% of totalr.m.s. voltage between the live conductors for thesum of the 2nd to 5th harmonic, or 12% max oftotal r.m.s. voltage between the live conductorsfor the sum of the 6th to 30th harmonic

Voltage interruption Supply must not be interrupted or at zero voltagefor more than 3ms at any time in the supply cycleand there should be more than 1 second betweensuccessive interruptions

Voltage dips Voltage dips must not exceed 20% of the peakvoltage or the supply for more than one cycle andthere should be more than 1 second betweensuccessive dips

‘capped’ by foam or filled with dried casting sand to create a noise, fire and vapourbarrier.

Large power cables may enter a cell via concrete trenches cast under the walland filled with dense dry sand below a floor plate. This method gives good soundinsulation, a fire barrier and it is relatively easy to add cables at a later date.

Electrical cabinet ventilation

Instrument errors caused by heat are particularly difficult to trace, as the instrumentwill probably be calibrated when cold; so they should be eliminated at the designand installation phases. Many instrumentation packages produce quite appreciablequantities of heat and if mounted low down within the confined space of a standard19-inch rack cabinet they may raise the temperature of apparatus mounted abovethem over the generally specified maximum of 40�C. Control and instrument cabinetsshould be well ventilated and it may be necessary to supplement individual ventilationfans by extraction fans high in the cabinet. Cabinet ventilation systems should havefiltered intakes that are regularly changed or cleaned.

Special attention should be given to heat insulation and ventilation when instru-ments are carried on an overhead boom. Thermal ‘tell-tale’ strips installed in cabinetsmay be considered as a good maintenance or fault finding device.


European safety standards and CE marking

The European Community has, since 1985, been developing regulations to achievetechnical harmonization and standards to permit free movement of goods within thecommunity. There are currently four directives of particular interest to the buildersand operators of engine test facilities (Table 10.4).

The use of the CE Mark (abbreviation for Conformité Europeen) implies that themanufacturer has complied with all directives issued by the EEC that are applicableto the product to which the mark is attached.

An engine test cell must be considered as the sum of many parts. Some of theseparts will be items under test that may not meet the requirements of the relevantdirectives. Some parts will be standard electrical products that are able to carrytheir individual CE marks, while other equipment may range from unique electronicmodules to assemblies of products from various manufacturers. The situation isfurther complicated by the way in which electronic devices may be interconnected.

If standard and tested looms join units belonging to a family of products, then thesum of the parts may comply with the relevant directive. If the interconnecting loomis unique to the particular plant, the sum of the CE-marked parts may not meet thestrict requirements of the directive.

It is therefore not sensible for a specification for an engine test facility to includean unqualified global requirement that the facility be CE marked. Some productsare specifically excluded from the regulations, while others are covered by theirown rules; for example the directive 72/245/EEC covers radio interference fromspark-ignition vehicle engines. Experimental and prototype engines may well fail tocomply with this directive: an example of the impossibility of making any unqualifiedcommitment to comply in all respects at all times with bureaucratic requirementsdrawn up by legislators unaware of our industry.

There are three levels of CE marking compliance that can be considered:

• All individual control and measuring instruments should individually complywith the relevant directives and bear a CE mark.

• ‘Standardized’ test bed configurations that consist exclusively of compliant instru-mentation, are configured in a documented configuration, installed to assembly

Table 10.4 EC directives for CE marking relevant to engine test facilities

Directive Reference Optional Mandatory

Electromagneticcompatibility

89/336/EEC 1 January 1996

Machinery 98/37EC 1 January 1993 1 January 1995Low voltage 73/23/EWG plus

93/68/EEC1 January 1995 1 January 1997

Pressure equipment November 1999 May 2002

212 Engine Testing

instructions/codes and have been subjected to a detailed and documented riskanalysis may be CE marked. This requires a test bed equipment supplier to defineand document such a package which would allow the whole ‘cell’ to be CEmarked.

• Project specific test cells. As stated above, the CE marking of the completehybrid cell at best will require a great deal of work in documentation and at worstmay be impossible, particularly if required retrospectively for cells containinginstrumentation of different generations and manufacturers.

The reader is advised to consult specific health and safety literature, or that produced

by trade associations, if in doubt regarding the way in which these directives should

be treated.

Note on ‘as built’ electrical documentation (drawing)standards

As stated at the start of this chapter, the electrical schematics of a test facility willbe used, more than other documentation, by technical staff having to modify andmaintain a test facility. It is therefore vital that such drawings reflect the ‘as built’state of the facility at the time of acceptance, and that they are arranged in a logicalmanner. There are many national and company standards covering the layout ofelectrical schematics and the symbols used. Some are significantly different fromothers, which can cause problems for system integrators and maintenance staff.Documentation standards should be stated in the functional specification (Chapter 1).

Safety interaction matrix

A key document of any integrated system is some form of safety or alarm interactionmatrix; it is the documentary proof that a comprehensive risk analysis has beencarried out. To be most effective, the base document should be verified between thesystem integrator and user group since it is the latter that have to decide, within aframework of safety rules, upon the secondary reactions triggered in the facility by aprimary event. The control logic of the building management system (BMS) has tobe integrated with that of the test control system; if the contractual responsibly forthe two systems is split then the task of producing an integrated safety matrix needsto be allocated and sponsored.

Summary

To enable a well-integrated system to be created, the electrical engineer will needto have both a complete operational specification and a mechanically completedfunctional specification. The final function and alarm matrix which defines how the


many component parts and subsystems must interact will be created as part of theelectrical design role.

It has been pointed out that electrical engineering regulation and practices aroundthe world differ and while electrical engineers may use the same words the localpractices may be significantly different, making integration of international projectsfraught with traps for the unwary.

Notation

Ampere, unit of electric current ABayonet cap BCAmbient temperature correction factor Ca

European Committee for Electro technicalStandardization

CENELEC

Cable grouping correction factor Cg

Thermal insulation correction factor Ci

Correction factor for the conductor operatingtemperature

Ct

Combined neutral and earth CNEPower factor (sinusoidal systems) cos 0Circuit protective conductor CPCCross-sectional area c.s.a.Overall cable diameter De

Electrical Contractors Association ECAEarthed equipotential bonding and automaticdisconnection

EEBAD

Earth leakage circuit breaker ELCBExtra-low voltage ELVElectromagnetic compatibility EMCElectromotive force e.m.f.Electromagnetic interference EMIEdison screw ESFrequency f

Functional extra-low voltage FELVGuidance note GNHigh breaking capacity (fuse) HBCHigh rupturing capacity (fuse) HRCHertz, unit of frequency HzSymbol for electric current I

Operating current (fuse or circuit breaker) I2Current to operate protective device IaDesign current IbFault current Id

214 Engine Testing

International Electrotechnical Commission IECInstitution of Electrical Engineers IEECurrent setting of protective device InTabulated current ItCurrent carrying capacity IzEarthing system ITKilo, one thousand times kKilovolt (1000V) kVLines of three-phase system L1, L2, L3Metre mMilli, one thousandth part of mMeg or mega, one million times MMilli-ampere mAMiniature circuit breaker MCBMoulded case circuit breaker MCCBMaximum demand MDMineral-insulated m.i.National Inspection Council for Electrical InstallationContracting

NICEIC

Potential difference p.d.Protective extra-low voltage PELVProtective earth PECombined protective earth and neutral PENProtective multiple earthing PMEProgrammable logic controller PLCProspective short-circuit current PSCPolyvinyl chloride p.v.c.Resistance (electrical) R

The total resistance of the earth electrode andthe protective conductor connecting it to exposedconductive parts

Ra

Resistance of the human body Rp

Residual current circuit breaker RCCBResidual current device RCDRoot-mean-square (effective value) r.m.s.Second, unit of time sConductor cross-sectional area S

Separated extra-low voltage SELVTime t

Earthing system (L1, L2, L3, PEN) TN-CEarthing system (L1, L2, L3, N, PE) TN-C-SEarthing system TN-SEarthing system TT


Symbol for voltage (alternative for V ) U

Alternating voltage Uac

Direct voltage Udc

Phase voltage Uo

Volt, unit of e.m.f. or p.d. VWatt, unit of power WReactance X

Impedance (electrical) Z

Earth loop impedance external to installation Ze

Earth fault loop impedance Zs

Phase angle �

Further reading

Electromagnetic Compatibility and Functional Safety. The Institution of Engineeringand Technology, www.theiet.org/factfiles.

Goedbloed, J.J. (1992) Electromagnetic Compatibility, ISBN: 0-13-249293-8, Pren-tice Hall, New Jersey.

Middleton, J. The Engineer’s EMC Workbook, ISBN: 0-9504941-3-5.Williams, T. (1992) EMC for Product Designers, ISBN: 0-7506-1264-9, Newnes,

Oxford.

11 Test cell control and dataacquisition

Introduction

This chapter deals with the test cell as an integrated working system and the practicalaspects of how that system is controlled, collects data and functions safely.

The forms of transducers used in engine test cells are described. Once again thetheme of system integration is important within this chapter as the switching on,normal, abnormal operation and shutting down of each piece of apparatus has to beconsidered for its possible effect on, or interaction with, any other apparatus andas part of the whole facility. The subject of data handling and post processing iscovered in Chapter 19.

Safety systems

Emergency stop function

The emergency stop (EM stop) function is a hard-wired system, normally quiteseparate from the alarm monitoring system embedded into the cell control system. Ingeneral practice it is a shutdown system initiated by humans rather then computers,although electrical links are often made to safety critical systems such as the firealarm system.

Most emergency stop systems may be visualized as a chain of switches in series allof which need to be closed before engine ignition or some other significant processcan be initiated. When the chain is opened, normally by use of manually operatedpress buttons, the process is shut down and the operated button is latched open untilmanually reset. This feature allows staff working on any part of the test cell complexto use the EM stop as a self-protection device by pressing an EM stop button andhanging a ‘Do not activate notice’ bearing name and time on it.

The traditional function of the emergency stop circuit is to shut off electricalpower from all devices directly associated with the running of the engine. In thismodel, dynamometers are usually switched to a ‘de-energized’ condition, fuel supply

Test cell control and data acquisition 217

isolation systems are activated, ventilation fans are shut down, but services sharedwith other cells are not interrupted and fire dampers are not closed.

Modern practice for cells fitted with electrical dynamometers is for the emergencystop function to shut down the engine and bring the shaft rotation to a stop via afast regenerative stop before cutting off power to the dynamometer. BS EN 60204recognizes the following categories:

• Uncontrolled stop (removal of power): Category 0. This is commonly usedin some engine cells fitted with manual control systems and two quadrantdynamometers.

• Controlled shutdown (fast stop with power removed at rotation stop or after timed

period): Category 1. This is the normal procedure for engine cells fitted withfour quadrant dynamometers where, under some circumstances, shutting downthe engine may not stop rotation.

• Controlled stop (fast stop with power retained at rotation stop): Category 2.Normally operated as a ‘fast stop’ rather than an EM stop in test cell practice(see below).

The control desk may be fitted with the following ‘emergency action’ buttons:

1. large red emergency stop button (latched) shutting down engine and cell systems;2. fast stop button (hinged cover to prevent accidental operation) operating only on

engine and dynamometer systems to stop rotation;3. fire suppression release button (wall-mounted and possibly break-glass pro-

tected).

It should be clear to all readers that operational training is required to be able todistinguish between situations requiring use of emergency control devices.

It is important to use the safety interaction action matrix, introduced in Chapter 10,to record what equipment will be switched to what state in the event of any of themany alarm or shutdown states being monitored.

Everything possible should be done to avoid spurious or casual operation of theEM stop system or any other system that initiates cutting off power to cell systems.Careful consideration has to be given before allowing the shutting down of plant thatrequires long stabilization times, such as humidity control for combustion air or oiltemperature control heaters. It is also important not to exceed the numbers of startsper hour of motors driving large fans, etc., that would cause overheating of windingsand the operation of thermal protection devices.

Operational and safety instrumentation

This term covers the basic instruments and controls necessary for the safe runningof the engine and cell equipment; unlike the EM stop the shutdown parameters aremonitored by the control system. Operational instrumentation safety systems can be

218 Engine Testing

either software operated, or hard wired,∗ with duplicate signals going to any associatedcomputer, and includes devices associated with the cell shutdown procedure, suchas fire dampers. The following indications would normally be included under thisheading:

• engine speed (shaft line speed too high);• oil pressure (after start-up, pressure too low);• facility cooling water flow and compressed air pressure;• signals associated with fire suppression system being available;• status of cell ventilation, purge system and fire dampers;• overspeed trips on both engine and dynamometer to deal with shaft failures;• cell door and shaft guard interlocks;• in-cell beacon/alarm automatically triggered by starting sequence.

Computerized monitoring of alarm signals

Primary alarms should refer to limits which, if exceeded, would endanger the facilitypersonnel and/or equipment. These are additional to alarms that are specific to theengine (which are often lower than the absolute limits of the facility or instrumenta-tion).

During a test, the condition of both engine and facility may be monitored con-tinuously by the test cell control computer so that the appropriate preprogrammedaction may be triggered should one or more parameters fall outside a preset limit.

Alarms are of two kinds: those safeguarding the equipment and those safeguardingthe integrity of the test. The first type will unconditionally cause protective actionsto be taken.

Most engine test computer software make available four levels of alarm for eachdata channel:

1. High level, warning2. High level, test shutdown3. Low level, warning4. Low level, test shutdown.

Such comprehensive warning arrangements are likely to be necessary on only a fewchannels; in most cases they will either be ‘switched off’ or assigned values outsidethe operating range. Some software suites allow stage-specific alarms to be set; thiscan be useful when, for example, particularly close control of fuel temperature isrequired during a consumption measurement, the warning limits can be closed onlyfor the critical period.

∗ The term ‘hard wired’, as used in this book, means that the instrument or other devicereceives and sends its status signal via its own unique cable rather than by way of a computer(software-dependent) signal or a multiplexing system. Increasing use of PLC-based buildingmanagement systems is making the concept of hardwiring of service status signals moredifficult.


Most engine test software contains a rolling buffer that allows examination of thevalues of all monitored channels for a short time before any alarm shutdown. Knownas ‘historic’ or ‘dying seconds’ logging, this buffer will also record the order inwhich the alarm channels have triggered, information of help in tracing the locationof the prime cause of a failure.

Security of supply

The possible consequences of a failure of the electricity supply, or of such factors asundetected fuse failures, must be carefully thought out. Particular dangers can ariseif the power is restored unexpectedly after a mains supply failure.

It is usual to design safety systems as normally de-energized (energize for normaloperation), but all aspects of cell operation must be considered. In many cases the cellcontrol logic should require that equipment, such as solenoid-operated fuel valves atthe cell entry, be reset after mains failure via an orderly, operator initiated, restartprocedure.

The installation of an uninterrupted power supply (UPS) for critical control anddata acquisition devices is to be recommended. This relatively inexpensive piece ofequipment need only provide power for the few minutes required to save or transfercomputer data and bring instrumentation to a safe state.

Building management systems (BMS) and services status displays

Test cell function is usually totally dependent on the operation of a number ofbuilding and site services, remote and hidden from the operator. A failure of one ofthese services will directly or indirectly (through alarm state caused by secondaryeffects) result in a cell shutdown. It is strongly recommended that indication of thestatus of each building service is available to the operator in the cell or, in the caseof multicell installations, at some common point. The common and time-wastingalternative is for operators to tour the facility in order to check the status of individualpieces of plant at start-up, a fault condition and shutdown. A centralized control andindication panel is of great value; however, such system status displays must indicatethe actual rather than the requested status.∗

A modern PLC-controlled BMS can have an interactive mimic display showingnot only the running state of services but also the status of fire dampers, valves andthe temperature and pressure of fluids. These PLC systems may have the services

∗ The Three Mile Island nuclear power station accident, initiated by a failure of the main feedwater pumps, was greatly exacerbated by incorrect actions taken by operations staff whoseonly indications showed a critical valve as being closed, but in fact this only representedthat the signal to close the valve was sent, in fact it was open. A ‘positive feedback’ lampin the control room indicating the true position of the valve had been eliminated in theoriginal construction to save time and money.

220 Engine Testing

system start, stop and alarm state logic programmed into them but the cell designersmust carefully consider whether each piece of plant’s primary control should comefrom the cell controller or the BMS; in the case of compressed air supply the choiceis clear but the case of fuel supply to the cell it may not be.

Test cell start-up and shutdown procedures

Failure of any test cell service, or failure to bring the services on stream in thecorrect order, can have serious consequences and it is essential to devise a logicalsystem for start-up and shutdown, with all the operator procedures, the necessaryalarms and automatic shutdowns. In all but highly automated production tests, avisual inspection of the unit under test must be made before each and every testrun. Long-established test procedures can give rise to complacency and lead toaccidents.

Checks before start-up

It will be assumed that routine calibration procedures have been followed, and thatinstrument calibrations are correct and valid; then the following type of checks shouldbe made or status confirmed:

• No ‘work in progress’ or maintenance labels are attached to any systemswitches.

• Engine/dynamometer alignment within set limits and each shaft bolt has beentightened to correct and recorded torque.

• Shaft-guard is in place and centred so that no contact with shaft is possi-ble (if appropriate, it is a good practice to rock the engine on its mountsto see that the rigging system, including exhaust tubing, is secure and flexescorrectly).

• All loose tools, bolts, etc. are removed from the test bed.• Engine support system is tightened down.• Fuel system is connected and leak proof.• Engine oil is at the correct level and oil pressure alarm hard wired into control

system.• All fluid services such as dynamometer water are on.• Fire system is primed.• Ventilation system is available and switched on to purge cell of flammable vapour

prior to start-up of engine.• Check that cell access doors remote from the control desk have warnings set to

deter casual entry during test.

Engine start process is initiated.


Checks immediately after start-up

• Oil pressure is above trip setting. In case of manually operated cells the alarmover-ride is released when pressure is achieved.

• If stable idle is possible and access around the engine is safe, a quick in-cellinspection should be made. Particularly look for engine fuel or oil leaks, cables orpipes chafing or being blown against the exhaust system and listen for abnormalnoises.

• In the case of a run scheduled for prolonged unmanned running, test the emer-gency shutdown system by operating any one of the stations.

• Restart and run test.

Checks immediately after shutdown

• Allow cooling period with services and ventilation left on.• Shut off fuel system as dictated by site regulations (draining down of day-tanks,

etc.).• Carry out data saving, transmission or back-up procedures.

The test sequence: modes of control

It should be recognized that a test programme for an engine coupled to a dynamome-ter is, first and foremost, a sequence of desired values of engine torque andspeed.

This sequence is achieved by manipulating only two controls: the engine speed(‘throttle’) and the dynamometer torque setting. (It is convenient to refer to thethrottle, even when it is an electric signal direct to the engine control unit or, in thecase of a diesel engine actuation of the fuel pump rack setting.)

The dynamometer torque is set according to the design of dynamometer used (seeChapter 8 for details).

For any given setting of the throttle, the engine has its own inherent torque-speedcharacteristic and similarly each dynamometer has its own torque-speed curve fora given control setting. The interaction of these two characteristics determines theinherent stability of the engine dynamometer combination.

Note on mode nomenclature

In this book, the naming convention of modes of control used in the UK andwidely throughout the English-speaking world has been adopted, the alternativenomenclature and word order used in some major German-speaking organizations isshown in Table 11.1.

222 Engine Testing

Table 11.1 Control modes nomenclatures

English (first term refers to

engine)

German (first term may refer

to dynamometer)

Speed nPosition AlphaTorque Torque

The engine or throttle control may be manipulated in three different ways:

• to maintain a constant throttle opening (position mode);• to maintain a constant speed (speed mode);• to maintain a constant torque (torque mode).

The dynamometer control may be manipulated:

• to maintain a constant control setting (position mode);• to maintain a constant speed (speed mode);• to maintain a constant torque (torque mode);• to reproduce a particular torque–speed characteristic (power law mode).

Various combined modes, usually described in terms of engine mode/dynamometermode, are possible.

Position/position mode

This describes the classic manually operated engine test. The throttle is set toa fixed position, the dynamometer control similarly set, and the system settlesdown, hopefully in a stable state. There is no feedback: this is an open loop sys-tem. Figure 11.1a shows a typical combination in which the engine has a fairlyflat torque–speed characteristic at fixed throttle opening, while the dynamometertorque rises rapidly with speed, typical of most but not all water-brake machines.The two characteristics meet at an angle approaching 90�, and operation is quitestable.

Certain designs of variable fill dynamometer with simple outlet valve control, seeChapter 8, may become unstable at light loads, this can lead to hunting, or to theengine running away. The two characteristics meet at an acute angle, Fig. 11.1b. Afriction brake, Fig. 11.1c operating at a given control setting, develops a torque almostindependent of speed and is clearly unsuitable for loading an engine in position mode.

Position/position mode is useful in system fault finding in that the inherentstability of the dynamometer, independent of control system influence may bechecked.


Dynamometer

Brake

DynamometerEngine

Torq

ue

Torq

ue

Torq

ue

Engine

Engine

Speed

(a)

(c)

(b)

Speed

DynamometerTorq

ue

Engine

(d)

SpeedSpeed

Figure 11.1 Control modes, engines and dynamometers: (a) positionmode stable hydraulic dynamometer; (b) position mode unstable hydraulicdynamometer; (c) position mode, friction brake; (d) position/speed mode; (e)position/torque, governed engine; (f) speed/torque mode; (g) torque/speedmode

Position and power law mode

This is a variation on position mode, in which the dynamometer controller is manip-ulated to give a torque–speed characteristic of the form

Brake torque= constant× speedn

When n= 2, this approximates to the torque characteristic of a marine propeller andthe mode is thus useful when testing marine engines. It is also a safe mode, tendingto prevent the engine from running away if the throttle is opened by an operator inthe cell.

224 Engine Testing

Dynamometer

Engine Engine

Engine

Torq

ue

Speed

Dynamometer

Torq

ue

Speed

Dynamometer

Torq

ue

Speed

(e) (f)

(g)

Figure 11.1 (Cont.)

Position and speed mode

In this mode, the throttle position continues to be set manually, but the dynamometeris equipped with an automatic controller which adjusts the torque absorbed by themachine to maintain the engine speed constant whatever the throttle position andpower output, Fig. 11.1d. This is a very stable mode, and is generally used forplotting engine torque–speed curves at full and part throttle opening.

Position and torque mode (governed engines)

Governed engines have a built-in torque–speed characteristic, usually slightly ‘droop-ing’ (speed falling as torque increases). They are therefore not suited for couplingto a dynamometer in speed mode. They can, however, be run with a dynamometerin torque mode. In this mode the automatic controller on the dynamometer adjusts


the torque absorbed by the machine to a desired value, Fig. 11.1e. Control is quitestable. Care must be taken not to set the dynamometer controller to a torque thatmay stall the engine.

Speed and torque mode

This is a useful mode for running in a new engine, when it is essential not to applytoo much load. As the internal friction of the engine decreases, it tends to developmore power and since the torque is held constant by the dynamometer the tendencyis for the speed to increase. This is sensed by the engine speed controller, which actsto close the throttle, Fig. 11.1f.

Torque and speed mode

This can be useful as an approximate simulation of the performance of a vehicleengine when climbing a hill. The dynamometer holds the speed constant while theengine controller progressively opens the throttle to increase torque at the set speed,Fig. 11.1g.

Precautions in modes of the four-quadrant dynamometer

Where the dynamometer is also capable of generating torque, it will be necessary totake precautions in applying some of the above modes of control. This applies whenthe dynamometer is in speed mode, Figs 11.1d, g. When running in either of thesemodes, the four-quadrant dynamometer may be controlled by logic that will maintainthe set speed even if the engine is being ‘motored’. This could clearly be dangerousshould the engine have failed in some way and has ceased to deliver power.

Throttle actuation

Except in the cases when an engine is controlled directly with a marine type ofcable system and lever at the control station, engines fitted with a speed controllever require the test cell to be equipped with some form of actuation that can beremotely and precisely operated, under manual or automatic control from the controlroom.

The actuators may be based on rotary servomotors, printed circuit motors or linearactuators, in all cases fitted with some form of positional feedback to allow closedloop control. Most actuators will be fitted with some form of stroke adjustmentto allow for 0 (shut)–100 per cent (WOT) positions to be determined for different

226 Engine Testing

engines and linkage configurations and some form of force limiter device. Typicalspecification ranges of proprietary devices are:

Constant force applied at Bowden cable connection 100–400NMaximum stroke length 100–160mmCable connector speed of travel 0.5–1.5m/sPositional repeat accuracy ±0.05mm

It must be remembered that the forces applied by the actuator to the engine lever

will depend on the mechanical linkage within the engine rigging.

It should be clear from the specification of positional repeat accuracy (above) thatcable and linkage used must be as free from backlash as is possible if good closedloop control is to be achieved. An important safety feature of a throttle actuator isthat it should automatically move the engine control lever to the ‘stop’ position inthe event of a power failure.

Many modern engines do not have a speed control lever, but instead require anelectrical signal proportional to speed requested (fly-by-wire). In the test cell thiscan be achieved using one of the following strategies:

1. As part of the engine rigging where part or all of a vehicle accelerator pedalmechanism is connected via cable or lever to a throttle actuator.

2. The electronic signal required is supplied via the test cell control computer as adiscrete ‘analogue out’ signal.

Problems in achieving control

An engine test system is an assembly of an engine, a dynamometer and variousactuators and peripherals. Each of these has its own control characteristics and, inmany cases, its own controller. These controllers have not been designed as a groupand it is not surprising that the combined control characteristics are often very farfrom ideal. The problem is compounded when the entire system is subjected to overallcomputerized control. At the simplest level, a dynamometer with a control loopintended to produce a particular torque–speed characteristic can generate instabilitiesquite absent in an old-fashioned manually controlled (brake) dynamometer. Thisgives a clue to investigating control instabilities: eliminate as much of the controlsystem as possible by switching to open-loop control.

It will also be found helpful, if two controllers are ‘fighting’ each other, to ensurethat they have widely differing time constants. In the common case of a speed-controlled engine coupled to a torque-controlled dynamometer, it is preferable thatthe latter should have a shorter response time than the former.

Some engines are inherently difficult to control because of the shape of theirtorque–speed characteristics. Turbocharged engines may have abrupt changes in the


slope of the power–speed curve, plus sluggish response due to the time taken forspeed changes in the turbocharger, which can make the optimization of the controlsystem very difficult.

Inexperienced attempts to adjust a full three-term PID (proportional/integral/dif-ferential) controller can lead to problems and it is well to record the settings beforestarting to make adjustments so that one can at least return to the starting point. It isgood practice to make sure that elementary sources of trouble, such as backlash and‘stick-slip’ in control linkages, are eliminated before delving into control variablesin software.

Choosing test control software

The world’s engine test industry is served by a relatively small number of companieswhose flagship product is a comprehensive suite of software designed specificallyfor controlling engine tests, acquiring test data and displaying results.

Several of these companies also make the major modules of instrumentationrequired for testing engines, such as dynamometers, engine fluid temperature con-trollers, fuel meters and exhaust emission equipment. However, many facilities aremade up of modules of plant coming from several of these specialist suppliers inaddition to locally made building services.

Such major software suites are the outcome of many man-years of work; it isnot recommended that even a highly specialized user, however experienced, shouldundertake the task of producing their own test control software from scratch.

Prospective buyers of automated test systems should if possible inspect com-petitive systems at the sites of other users doing similar work. As with any com-plex system, the following general questions should be considered in no order ofimportance:

• How intuitive and supportive to the work does the software appear to be; or doesthe software provider understand the world in which the software has to function?

• Purchase cost is clearly important, but also consider the cost of ownership includ-ing technical support, upgrades and additional licences.

• Can the post-processing and reporting required be carried out directly by the newsoftware or can data be transferred to other packages used on site?

• How good is the training and service support in the user’s geographical location?• In the case of international companies, does the software support multilanguage

display of screen displays?• Has the system been successfully used in the user’s industrial sector?• Can existing instrumentation or new third party instrumentation be used with the

system without performance degradation?• Is data security robust and can the post-processing, display and archiving required

be accommodated?• Can the data generated by the new system be stored in the existing data format?

228 Engine Testing

Test cell computer roles and connected devices

This section is based generally on the assumption that each cell will have its individualPC in a hierarchical position above any of the devices described below. The subjectof post-test data processing is covered in Chapter 19.

Editing and control of complex engine test sequences

The typical modern engine test programme consists of a long succession of stagesof speed and torque settings together with the demand values sent to other deviceswithin the test cell system. Together these form a complete test profile or ‘sequence’.It is beyond the capacity of a human operator to run such test sequences accuratelyand repeatedly.

Although the terminology used by test equipment suppliers may vary, the prin-ciples involved in building up test profiles will be based on the same essentialcomponent instructions.

Test sequence ‘editors’ range from those that require the operator to have knowl-edge of software code to those that present the operator with an interactive ‘formfill’ screen. The nature and content of the questions shown on the screen will dependon the underlying logic which determines how the answers are interpreted. It isimportant that the user should understand the interaction logic of the test cell logic,otherwise there is a risk of calling up combinations of control that the particularsystem cannot run, though it may not be able to indicate where the error lies. Forexample, if the control mode is set at throttle: position and dynamometer: speed, theeditor will ‘expect’ instructions in terms of these parameters; it will not in this casebe able to accept a throttle control instruction in terms of torque when it ‘needs’percentage open.

Each sequence will be made up of a series of stages, each either steady state,engine speed and torque constant, or transient, covering a move from one setting ofspeed and torque to another within a specified time or ‘ramp rate’. Of course speedand torque may also remain constant and other individually controllable parametersmay be varied.

Test sequence elements

In most cases, the test sequence editors embodied in a suite of software designed forthe control of engine test beds are based on a simple ‘form fill’ layout. The followingelements are involved:

• Mode of control. Some older analogue-based control systems require engine speedand load to be brought down to some minimum value between each stage, butmodern digital controllers should be able to make a ‘bumpless transfer’ betweenmodes of control.


• Engine speed and torque or throttle position. The way in which these parametersare set will depend on the control mode. A good sequence editor will presentonly viable options.

• Ramp rate. This is the acceleration required or the time specified for transitionfrom one state to the next.

• Duration or ‘end condition’ of stage. The duration of the stage may be definedin several ways:

– at a fixed time after the beginning of the stage;– on a chosen parameter reaching a specified value;– on reaching a specified logic condition, e.g. on completion of a fuel con-

sumption measurement.

• Choice of next stage. At the completion of each stage (stage x), the editor will‘choose’ the next stage. Typical instructions governing the choice are:

– run stage x+1;– rerun stage x a total of y times (possible combinations of ‘looped’ stages may

be quite complex and include ‘nested’ loops);– choose next stage on basis of a particular analogue or digital state being

registered (conditional stage).

• Events to take place during a stage. Examples are the triggering of ancillaryevents, such as fuel consumption measurement or smoke density measurement.These may be programmed in the same way as duration or end condition, above.

• Nominated alarm table. This is a set of alarm channels that are activated duringthe stage in which they are ‘called’. The software and wired logic must preventsuch programming from overriding ‘global’ safety alarms.

Data acquisition and the transducer chain

In this section, we consider data collection from single transducers sensing pressureor temperature; torque and speed are covered in Chapter 8 and data from complexinstruments is covered later in this chapter.

The information defining the state of any measured parameter travels via cablefrom the transducer, through signal conditioning, to a storage device and one or moredisplays. During this journey it is subject to a number of possible sources of errorand corruption covered in Chapter 10.

A well-integrated system will have every detail of the chain specified to suit therequirements and inherent characteristics of the parameter being measured.

Calibration of the signal chain

All professional engine test software suites will contain calibration routines for allcommon transducers and instruments. Such routines will always allow fixing of zeroand full span points and compensation for non-linearity of output.

230 Engine Testing

For the most demanding calibration the complete chain from transducer to finaldata store and indication needs to be included, but for normal work it is neitherpractical nor cost/time effective. Common practice is to use certified transducers andan electrical signal from a calibrated instrument which is injected into the measuringsystem at the point at which the transducer is plugged in. Several measuring pointsthroughout the transducer’s range are simulated and checked against the display.

Transducer boxes and distributed I/O

In order to measure the various temperatures and pressures of the UUT (engine orvehicle), transducer probes have to be attached and their signal cables connected intothe data acquisition system. In many cases the transducer probes will be intrusiveand fitted via test points sealed with compression fittings. In addition to the singleprobe cables, there will be cables from other instruments such as optical encodersrequiring connection within the test cell in order for the signals, in raw form, to betransmitted to the signal conditioning device.

Modern practice is to reduce the vulnerability of signal corruption by conditioningthe signal as near as possible to the transducer and to transmit them to the computervia a parallel SCSI or IEEE 1394 high speed serial bus; such implementations,generally known as distributed I/O, can greatly reduce complexity of installation.

Older installations will have discrete cables in a loom from the transducer box,through the cell wall to a connection strip in the control cabinet near the controldesk. This network of short transducer cables needs to be marshalled and collectedat a transducer box near the engine.

Transducer boxes will contain some or all of the following features:

• external, numbered or labelled, thermocouple, PRT or pressure line sockets;• external labelled sockets for dedicated channels or instruments;• power supply sockets for engine mounted devices, such as an ECU.

Transducer boxes may have upwards of 50 cables or pipes connected to them fromall points of the engine under test; this needs good housekeeping by clipping insublooms to prevent an unsightly, and difficult to use, tangle.

The transducer box can be positioned:

1. On a swinging boom positioned above the engine, which gives short cable lengthsand an easy path for cables between it and the control room. However, unlesspositioned with care it will be vulnerable to local overheating from radiation orconvection from the engine; in these cases, forced ventilation of the enclosedboom duct and transducer box should be provided.

2. On a pedestal alongside the engine but this may give problems in running cableback to the control room.

3. On the test cell wall with transducer cables taken from the engine via alightweight boom.


The choice of instruments and transducers

It is not the intention to attempt a critical study of the vast range of instrumentationthat may be of service in engine testing. The purpose is rather to draw the attentionof the reader to the range of choices available and to set down some of the factorsthat should be taken into account in making a choice.

Table 11.2 lists the various measurements. Within each category the methods arelisted in approximately ascending order of cost.

Note that:

• Many of the simpler instruments, e.g. spring balances, manometers, Bourdongauges, liquid-in-glass thermometers, cannot be integrated with data loggingsystems. This may or may not be important.

• Accuracy costs money. Most transducer manufacturers supply instruments to sev-eral levels of accuracy with steeply increasing cost. In a well-integrated systema common level is required; over- or underspecification of individual parts com-pared with the target standard either wastes money or compromises the overallperformance.

• It is cheaper to buy a stock item rather than to specify special features.• Some transducers, particularly force and pressure, can be destroyed by overload.

Ensure adequate capacity or overload protection.• Always read the maker’s catalogue with care, taking particular note of accuracies,

overload capacity and fatigue life.

Time intervals and speed

Time can be measured to accuracy greater than is necessary in most engine testing,but the precise location of events in time is very much more difficult.1 When linkedevents that occur very close together in time are sensed in different ways, withdifferent instruments, it can be very difficult to establish the exact order or to detectsimultaneity.

In general, a tachometer should only be used for speed measurement when thereading is not to be used as one component in calculating another quantity such aspower. It is better practice to count the revolutions.

A single impulse trigger, such as a hole in the rim of the flywheel, is satisfactoryfor counting purposes but not ideal for locating top dead centre (see Chapter 14 fordetailed coverage of t.d.c. measurement). Multiple impulses picked up from a starterring gear or flywheel drillings used by the engine’s ECU may be acceptable. Manydynamometers are fitted with a 60-tooth wheel and inductive pick-up as standard foruse by their own control system.

For precise indications of instantaneous speed and speed crank angle down totenths of a degree an optical encoder is necessary. Mounting arrangements of anencoder on the engine may be difficult and if, as is usually the case, drive is takenfrom the non-flywheel end of the crankshaft torsional effects in the crankshaft may

232 Engine Testing

Table 11.2 Common instrumentation and transducers for frequently requiredmeasurements

Measurement Principal applications Method

Time interval Rotational speed TachometerSingle impulse triggerStarter ring gearShaft encoder

Force,quasistatic

Dynamometer torque Dead weights and springbalance

Hydraulic load cellStrain gauge transducer

Force, cyclic Stress and bearing loadinvestigations

Strain gauge transducerPiezoelectric transducer

Pressure,quasistatic

Flow systems; lubricant, fuel,water, pressure charge,exhaust

Liquid manometerBourdon gaugeStrain gauge transducer

Pressure,cyclic

In-cylinder, inlet, exhaustevents. Fuel injection

Strain gauge transducerCapacitative transducerPiezoelectric transducer

Position Throttle and other controls Mechanical linkage andpointer, counter

LVDT transducerShaft encoderStepper motor

Displacement,cyclic

Valve lift, injection needle lift Inductive transducerHall effect transducerCapacitative transducer

Acceleration Engine balancing, NVH Strain gauge accelerometerPiezoelectric accelerometer

Temperature Cooling water, lubricant, inletair, exhaust, in-cylinder,mechanical components

Liquid-in-glassVapour pressureLiquid-in-steelThermocouplePRTThermistorElectrical resistanceOptical pyrometerSuction pyrometer


lead to errors which may differ from cylinder to cylinder (see Chapter 14 for adiscussion of the problems involved).

Force, quasistatic

The strain gauge transducer has become the almost universal method of measuringforces and pressures. The technology is very familiar and there are many sourcesof supply. There are advantages in having the associated amplifier integral with thetransducer as the transmitted signals are less liable to corruption, but the operatingtemperature may be limited.

The combination of dead weights and spring balance is still quite satisfactory formany purposes.

Hydraulic load cells are simple and robust devices still fairly widely available,particularly used in portable dynamometers.

Force, cyclic

The strain gauge transducer has a limited fatigue life and this renders it unsuitablefor the measurement of forces that have a high degree of cyclic variation, though ingeneral there will be no difficulty in coping with moderate variations such as resultfrom torsional oscillations in drive systems.

Piezoelectric transducers are immune from fatigue effects but suffer from thelimitation that the piezoelectric crystal that forms the sensing element produces anelectrical charge that is proportional to pressure change and which thus requiresintegration by a charge amplifier. By definition, the transducer is unsuitable forsteady-state measurements but is effectively the universal choice for in-cylinder andfuel injection pressure measurements. Piezoelectric transducers and signal condi-tioning instrumentation are in general more expensive than the corresponding straingauge devices.

Measurement of pressure

To avoid ambiguity, it is important to specify the mode of measurement whenreferring to a pressure value. The three modes are (Fig. 11.2):

• absolute;• relative;• differential.

Pressures measured on a scale which uses a complete vacuum (zero pressure as ina theoretical vessel containing no molecules) as the reference point are said to beabsolute pressures.

234 Engine Testing

Absolute zeropressure

Absolutepressures

Differentialpressures

Atmosphericpressure

Relative orgauge

pressures

Figure 11.2 Diagram showing relationship between absolute, relative and dif-ferential pressures

Atmospheric pressure∗ at the surface of the earth varies, but is approximately105 Pa (1000mbar). This is an absolute pressure because it is expressed with respectto zero pressure. Most engine test instrumentation is designed to measure pressurevalues that are expressed with respect to atmospheric pressure, thus they indicate zerowhen their measurement is recording atmospheric pressure. In common parlance,relative pressures are referred to as gauge measurements. Because of the differencebetween an absolute pressure value and a gauge pressure, the variable value ofatmospheric pressure engine test facilities needs to have a barometric measurementrecorded as a reference in all tests.

Engine inlet manifold pressure will vary between pressures above and (generally)below the ‘reference’ barometric pressure. This is another form of relative pressuresometimes called a ‘negative gauge pressure’.

In applications such as the measurement of flow through an orifice, where it isnecessary to measure the difference in pressure between two places, the referencepressure may not necessarily be either zero or atmospheric pressure but one of themeasured values; such measurements are of differential pressures.

Electronic pressure transducers

There are large ranges of strain gauge and piezoelectric pressure transducers madefor use in engine testing, most having the following characteristics:

∗ For reference concerning the range of barometric pressures, the highest recorded atmosphericpressure, 108.6 kPa (1086 mbar) occurred in Mongolia, 19 December 2001. The lowestrecorded atmospheric pressure, 87.0 kPa (870 mbar), occurred in the Western Pacific duringTyphoon Tip on 12 October 1979.


• Pressure transducers normally consist of a metal cylinder made of a stainlesssteel, with a threaded connection at the sensing end and a signal cable attachedat the other.

• The transducer body will contain some form of device able to convert the pressuresensed at the device inlet into an electrical output. The signal may be analoguemillivolt, 4–20mAmp or a digital output via a CAN bus serial communicationsinterface.

• The output cable and connective circuitry will depend on the transducer typewhich may require 4, 3 or 2 wire connection one of which will be a stabilizedelectrical supply.

• Each transducer will have an operating range that has to be carefully chosen forthe individual pressure channel, they will often be destroyed by overpressurizationand be ineffective at measuring pressures below that range.

• Each transducer will be capable of dealing with specific pressurized media; there-fore, use with fuels or special gases must be checked at the time of specification.

Capacitative transducers, in which the deflection of a diaphragm under pressure issensed as change in capacity of an electrical condenser, may be the appropriatechoice for low pressures, where a strain gauge or piezoelectric sensor may not besufficiently sensitive.

Calibration of the complete pressure channel may be achieved by use of a certified,portable calibrator while some special pressure transducers such as barometric sensorsand differential pressure transducers may need annual off-site certification.

Pressure sensors used in the engine testing environment, because of their expenseand vulnerability to damage, are normally mounted within the transducer box withflexible tubes connecting them via self-sealing couplings to their measuring point.

Other methods of pressure measurement

The liquid manometer has a good deal to recommend it as a device for indicatinglow pressures. It is cheap, effectively self-calibrating and can give a good indicationof the degree of unsteadiness arising from such factors as turbulence in a gas flow.It is recommended that manometers are mounted in the test cell with the indicatingcolumn fixed against the side of the control room window. Precautions regarding theuse of mercury have reduced its common use in engine testing from common to rare.

The traditional Bourdon gauge with analogue indicator is the automatic choicefor many fluid service sensing points, such as compressed air receivers.

Displacement

Inductive transducers are used for ‘non-contact’ situations, in which displacementis measured as a function of the variable impedance between a sensor coil and amoving conductive target.

236 Engine Testing

The Hall effect transducer is another non-contact device in which the movementof a permanent magnet gives rise to an induced voltage in a sensor made from goldfoil. It has the advantage of very small size and is used particularly as an injectorneedle lift indicator.

Capacitative transducers are useful for measuring liquid levels, very small clear-ances or changes due to wear.

Acceleration/vibration

There are two basic types of accelerometer, based respectively on strain gauge andpiezoelectric sensors. Piezoelectric sensors should not be used for frequencies of lessthan about 3Hz, but tend to be more robust than strain gauge units.

Temperature measurement – thermocouples

Most of the temperatures measured during engine testing do not vary significantly ata high rate of change, nor is the highest degree of accuracy cost effective. The typesof thermocouple commonly used in engine testing are shown in Table 11.3.

For the majority of temperature channels, the most commonly used transducer isthe type K thermocouple, having a stainless steel grounded probe and fitted with itsown length of special thermocouple cable and standard plug. To produce a calibratedtemperature reading, all thermocouples require to be fitted within a system havinga ‘cold junction’ of known temperature with which there own output is compared.Thermocouples of any type may be purchased with a variety of probe and cablespecifications depending on the nature of their installation; they should be consideredas consumable items and spares kept in stock.

Temperature measurement – PRTs

When accuracy and long-term consistency in temperature measurement higher thanthat of thermocouples, or for measuring temperatures below 0�C, is required, thenplatinum resistance thermometers (PRT) are to be recommended.

The PRT works on the principle of resistance through a fine platinum wire as afunction of temperature. PRTs are used over the temperature range−200�C to 750 �C.

Table 11.3 Commonly used thermocouple types

Type Internal materials Temperature range

Type T Cu-Ni/Cu 0�C to +350�CType J Fe-Ni/Cu 0�C to +800�CType K Ni/Cr-Ni/Al/Mn −40�C to +1200�C


Because of their greater durability or accuracy over time, it is common practice touse PRTs in preference to thermocouples within device control circuits. They requiredifferent signal conditioning from thermocouples and, for engine rigged channels,are fitted with dedicated plugs and sockets within the transducer box.

Thermistors

Thermistors (thermally sensitive resistors) have the characteristic that, dependent onthe nature of the sensing ‘bead’, they exhibit a large change in resistance, eitherincreasing or decreasing, over a narrow range of temperature increase. Because theyare small and can be incorporated within motor windings, they are particularly usefulas safety devices. Thermistors are generally available in bead or surface mountingform and in the range from −50�C to 200�C.

Pyrometers

Optical or radiation pyrometers, of which various types exist, are non-contact temper-ature measuring devices, used for such purposes as flame temperature measurementand for specialized research purposes, such as the measurement of piston ring surfacetemperatures by sighting through a hole in the cylinder wall. They are effectively theonly means of measuring very high temperatures.

Finally, suction pyrometers, which usually incorporate a thermocouple astemperature-sensing device, are the most accurate available means of measuringexhaust gas temperatures,

Other temperature-sensing devices

Once a standard tool in engine testing, liquid-in-glass thermometers are now veryrarely seen; however, they are cheap, simple and easily portable. On the other hand,they are fragile, not easy to read and have a relatively large heat capacity and a slowresponse rate. They retain some practical use during services commissioning andsensing of plant in remote locations. The interface between the body, solid, liquidor gas, of which the temperature is to be measured and the thermometer bulb needscareful consideration.

Vapour pressure and liquid-in-steel thermometers present the same interface prob-lems are not as accurate as high-grade liquid-in-glass instruments, but are moresuitable for the test cell environment and are more easily read.

Mobile, hand-held thermal imaging devices are more useful for fault finding andsafety work than for collection of test data other than during installed on-site workwith large stationary engines.

Like all other aspects of temperature measurement, there is much guidance avail-able in the literature and on the internet.2

238 Engine Testing

Smart instrumentation and devices

The test cell control and data acquisition computer will, in addition to taking inindividual transducer signals, also have to switch on and off and acquire data fromcomplex modules of instrumentation, a list of which includes

• fuel consumption devices: gravimetric or mass flow;• oil consumption;• engine ‘blow-by’;• exhaust gas emission analysis equipment;• combustion analysis (indicating) instrument.

The integration of these devices within the control system will require special soft-ware drivers to allow for communication covering basic control, data acquisition andcalibration routines.

The interface between device and test cell computer may be by way of con-ventional analogue or digital I/O, by serial interface such as RS232 or IEEE or,increasingly, through a local area network (LAN) (see Chapter 19).

Computerized calibration procedures

Software routines will be provided to allow the operator to calibrate the variousmeasuring devices and systems. Often these routines follow a ‘form fill’ format thatrequires the operator to type in confirmation of test signals in the correct order,permitting zero, span and intermediate and ambient values to be inserted in thecalibration calculations. Such software should not only lead the operator throughthe calibration routines, but should also store and print out the results in such away as to meet basic quality control and certification procedures. The appropriatelinearization and conversion procedures required to turn transducer signals into thecorrect engineering units will be built into the software.

Control for endurance testing and ‘unmanned’ running

These tests are very expensive to run and call for a high level of performance andreliability in the control and data logging systems. Duplication of transducers andcareful running of cables to prevent deterioration by wear or heat over long periodsshould be considered. The amount of data generated over 1000 hours of cyclicrunning can be enormous, so data filtering and compression are usually required.

Control for thermal shock and thermal cycling tests

These call for special programming arrangements, ‘end condition’ of stage beingoften defined by the attainment of a given temperature. There will often be the need tooperate fluid control valves in strict sequence in order to direct engine coolant flows


to or from chilled water buffer stores. Since this type of testing may be expectedto induce a failure in some part of the engine system the failure monitoring systemmust be designed accordingly.

Control for transient or dynamic engine testing

For some purposes, such as the analysis of the performance of electronic enginemanagement systems (see also Chapter 19), it may be necessary to record data ata speed at which accurate ‘time stamping’ of individual events becomes difficult.‘Time skew’ can arise when the time taken to record all the values correspondingto an event taking place at a given instant is significant relative to the intervalbetween successive events. High speed data recording can produce unwanted detail,for instance cyclic variations in crankshaft speed when rotational speed is sensed onthe basis of very short measuring intervals.

Any vehicle user is aware that the performance of an engine is to a considerableextent judged by its characteristics when changing state. A cloud of black smokefrom a truck pulling away from traffic lights or hesitant throttle response whenaccelerating are just two of many examples of inadequate performance that is a directconsequence of unsatisfactory transient behaviour on the part of the engine.

It is perhaps worth pointing out that while, in what may be described as ‘classical’steady state testing, the transitions between successive test conditions are unimpor-tant, in transient testing the transient states are the test sequences.

The terms transient or dynamic testing are somewhat subjective terms and covertesting that is, in practice, almost confined to automotive engines, since industrialand marine engines are rarely subject to the almost continuous variations in load andspeed that are characteristic of the automotive engine environment. Test sequencesthat are neither transient nor dynamic are usually referred to as ‘steady state’ andalthough that is scientifically incorrect, the term covers sequences where the engine istaken to a set condition and held at that operating point for a finite time, normally inexcess of 5 seconds, but in some cases any period above 100ms, while measurementsare taken.

Up to the 1990s, the majority of formalized transient engine testing was associatedwith emissions testing carried out in accordance with the kind of test proceduredescribed in Chapter 16.

Since that time the work associated with engine and drive train calibration ormapping and track simulation in the motorsport industry has greatly increased, as hasthe sophistication of the models, which may be required to simulate both the roadload and the vehicle power train; such test sequences can be considered as dynamic.Importantly dynamometer technology has provided the tools that allow for speed andtorque changes in the test-bed drive line to match those experienced on the road andtrack to ever better resolution.

What may be considered a highly dynamic test sequence in one industrial sectormay be considered positively pedestrian in another. Therefore, in this chapter transientand dynamic are classified in terms of four characteristic time scales where the times

240 Engine Testing

quoted represent the time in which the control instruction is sent and the change inforce at the a.c. dynamometer air gap is sensed:

• <2ms is ‘high dynamic’;• <10ms is dynamic;• <50ms is transient;• >100ms is steady state not so much defined by the period at with the engine

state is held but also because the data are required at an operation point ratherthan during the transition between points.

Under these classifications falls the wide range of work required of engine test cells.The test engineer involved in transient testing, which is not controlled by self-

adjusting, boundary-seeking software tools of the type discussed in Chapter 19, isfaced with two challenge: firstly to ensure that the test sequence accurately modelsthe conditions experienced by the engine in service and secondly to ensure that thesequence is precisely repeatable.

The favoured modern tool is the use of a direct coupled four-quadrant, a.c.dynamometer with computer control; such a machine can be designed to react suffi-ciently rapidly to simulate all the transient conditions experienced by the engine onthe road.

In the range up to 0.2 s, we are concerned essentially with torsional vibrations inthe engine–transmission–road wheel complex ranging from two-mass engine–vehiclejudder, with a frequency typically in the range 5–10Hz, to much higher frequencyoscillations involving the various components of the power train. Subtle effects canoccur as a result of interaction with the engine mountings. Rotation of the engine onits mountings not only affects the dynamics of the whole system but may disturb thethrottle linkage, with consequent variations in the throttle position.

This kind of power train investigation, however, involves the engine to a secondarydegree; in fact, for such work the engine may now be replaced by a lower inertiaelectric motor. These tests are mainly concerned with the development of transmissionsystems and are thus rather outside the scope of this book.

In the transient test time scale work is characteristic of gear shifts; correctbehaviour in this range is critical in a system called upon to simulate these events.This is a particularly demanding area, as the profile of gearshifts is extremely vari-able. At one extreme we have the fast automatic changes of a race car gearbox, at theother gear changes in the older commercial vehicle, which may take more than 2 s,and in between the whole range of individual driver characteristics, from aggressiveto timid.

Whatever the required profile, the dynamometer must impose zero torque onthe engine during the period of clutch disengagement and this requires precisefollowing of the ‘free’ engine speed. The accelerating/decelerating torque requiredis proportional to dynamometer inertia and the rate of change of speed:

T = Id�/dt


where T = torque required in Nm, I = dynamometer inertia in kg m2, and d�/dt =rate of speed change in rad/s. This requirement is a major factor in the choice ofdynamometers for transient testing (see Chapter 8).

A gear change involves throttle closing, an engine speed change, up or down,of perhaps 2000 rev/min, and the reapplication of power. Ideally, at every instantduring this process the rate of supply of fuel and air to the engine cylinders and theinjection and ignition timings should be identical with those corresponding to theoptimized steady state values for the engine load and speed at that instant; it will beclear that this optimizing process makes immense demands on the mapping of theengine management system.

A special problem of engine control in this area concerns the response of aturbocharged engine to a sudden demand for more power. An analogous problemhas been familiar for many years to process control engineers dealing with themanagement of boilers. If there is an increase in demand, the control system increasesthe air supply before increasing the fuel input; on a fall in demand, the fuel input isreduced before the air. The purpose of this ‘air first up–fuel first down’ system is toensure that there is never a deficiency of air for proper combustion.

In the case of a turbocharged engine only specialized means, such as variableturbine geometry, can increase the air flow in advance of the fuel flow. A reduction invalue from the (optimized) steady state value, with consequently increased emissions,is an inevitable concomitant to an increase in demand. The control of maximumfuel supply during acceleration is a difficult compromise between performance anddriveability on the one hand and a clean exhaust on the other.

References

1. BS 3403 Specification for Indicating Tachometer and Speedometer Systems for Industrial,

Railway and Marine Use.

2. BS 1041 Parts 1 to 7 Temperature Measurement.

Further reading

BS 6174 Specification for Differential Pressure Transmitters with Electrical Outputs.National Physical Laboratory. Good Practice Guides. http://www.npl.co.uk/.

12 Measurement of fuel, combustionair and oil consumption

Introduction

In this chapter a review is made of the instruments required to measure the consump-tion of the liquid fuel, combustion air and lubricating oils during engine running ina test bed. Until the advent of instruments capable of measuring ‘instant’ air andfluid flow, the standard methods recorded cumulative consumption over a period oftime, or measured the number of revolutions over which a known mass of liquid fuelwas consumed. While these methods and instruments are still viable, more and moretesters are required to record the actual transient consumption during test sequencesthat simulate ‘real-life’ driving conditions.

Modern engine designs incorporate a number of fuel spill-back or fuel returnstrategies that make the design of test cell consumption systems more complex thanbefore. In simple terms, it is necessary now to measure the amount of fuel enteringthe metering and conditioning system, rather than leaving it, for the engine. Thesystems frequently have to control the pressure of the fuel return line and removevapour bubbles and the heat energy picked up through the engine’s fuel rail andpressure regulator circuit. Liquid fuel conditioning and consumption measurementis no longer a matter of choosing the appropriately sized instrument, but rather ofdesigning a complex pipe circuit incorporating those instruments, heat exchangersand the engine.

Measurement of the air consumption of the engine is covered in principle andpractice.

Measurement instruments for liquid fuel consumption

For all liquid fuel consumption measurement, it is critical to control fuel temperaturewithin the metered system as far as is possible. Therefore, most modern cells havea closely integrated temperature control and measurement system. The condition ofthe fuel returned from the engine can cause significant problems as it may return atpulsing pressure, considerably warmer than the control temperature and containingvapour bubbles. This means that the volume and density of the fluid can be vari-able within the measured system, and that this variability has to be reduced as far

Measurement of fuel, combustion air and oil consumption 243

as is possible by the overall metering system and the level of measurement uncer-tainty taken into account (see Chapter 7, section Engine fuel temperature controlfor a full discussion concerning the effect of fuel conditioning on fuel consumptionmeasurement).

Cumulative flow meters

There are two generic types of fuel gauge intended for cumulative measurement ofliquid fuels on the market:

• Volumetric gauges, which measure the number of engine revolutions taken toconsume a known volume of fuel either from containment of known volume oras measurement of flow through a measuring device.

• Gravimetric gauges, which measure the number of engine revolutions taken toconsume a known mass of fuel.

Figure 12.1 shows a volumetric gauge in which an optical system gives a precisetime signal at the start and end of the emptying of one of a choice of calibrated

Vent to fuel tank

Sidetube

Sensors

Fuel supply

Spillback connection

To engine

Spacers

Calibratedvolumes

Figure 12.1 Volumetric fuel consumption gauge

244 Engine Testing

Valve

Diaphragm

Load

cell

Spillback

To engine

Fuel supply

Figure 12.2 Gravimetric, direct weighing, fuel gauge

volumes. This signal actuates a counter, giving a precise value for the number ofengine revolutions made during the consumption of the measured volume of fuel.

Figure 12.2 shows a gravimetric gauge designed to meter a mass rather than avolume of fuel, consisting essentially of a vessel mounted on a weighing cell fromwhich fuel is drawn by the engine. The signals are processed in the same way as forthe volumetric gauge. A typical specification being:

• measuring ranges: 0–150 kg/h, 0–200 kg/h, 0–360 kg/h;• fuels: petrol and diesel fuels, special versions for 10 per cent methanol or ethanol;• computer interface: serial: RS232C, 1–10V analogue plus digital I/O;• electrical supply required typically 0.5–2.5 kW for measuring and fuel

conditioning.

A further type of gravimetric fuel gauge, in which a cylindrical float is suspendedfrom a force transducer in a cylindrical vessel, exists. The change in flotation force


is then directly proportional to the change of mass of fuel in the vessel. This designof gauge has the particular advantage that it is insensitive to vertical accelerationsand is thus suitable for shipboard use.

These gravimetric gauges have the advantage over the volumetric type that themetered mass may be chosen at will, and a common measuring period may be chosen,independent of fuel consumption rate. The specific fuel consumption is deriveddirectly from only three measured quantities: the mass of fuel consumed, the numberof engine revolutions during the consumption of this mass and the mean torque. Thisis another reason for the inherently greater accuracy of cumulative fuel consumptionmeasurement; rate measurements involve four or, in the case of volumetric meters,five measured quantities.

All gauges of this type have to deal with the problem of fuel spillback fromfuel injection systems and Fig. 12.3 shows a circuit incorporating a gravimetric fuelgauge which deals with this matter. When a fuel consumption measurement is to bemade, a solenoid valve diverts the spillback flow, normally returned to the headertank during an engine test, into the bottom of the fuel gauge. It is not satisfactoryto return the spillback to a fuel filter downstream of the fuel gauge, since the airand vapour always present in the spilled fuel lead to variations in the fuel volumebetween fuel gauge and engine, and thus to incorrect values of fuel consumption.

Mass flow or consumption rate meters

There are many different designs of rate meter on the market and the choice of themost suitable unit for a given application is not easy. The following factors need tobe taken into account:

• volumetric or gravimetric;• absolute level of accuracy;• turn-down ratio;• sensitivity to temperature and fuel viscosity;• pressure difference required to operate;• wear resistance and tolerance of dirt and bubbles;• analogue or impulse-counting readout;• suitability for stationary/in-vehicle use.

Several designs of positive-displacement (volumetric) fuel gauges make use of afour-piston metering unit with the cylinders arranged radially around a single-throwcrankshaft. Crankshaft rotation is transmitted magnetically to a pulse output flowtransmitter. Cumulative flow quantity and instantaneous flow rate are indicated andthese meters are suitable for in-vehicle use. A high turn-down ratio is claimed and onedesign includes a pressure-sensing system to eliminate leakage errors. A disadvantageis the appreciable pressure drop, which may approach 1 bar, required to drive themetering unit. For large flow rates metering units employing meshing helical rotorsare usual.

246 Engine Testing

Vent line

Gravimetricfuel gauge

column

Injection pump

Spillbackdiverter valve

Measure/runsequencing

valve

Day tank

Figure 12.3 Spillback and associated valves in a gravimetric fuel gauge circuit

Some mass flow meters make use of the Coriolis effect, in which the fuel ispassed through a vibrating U-tube, Fig. 12.4.

Some mass flow meters are able to measure spillback, but with others it may benecessary to use two sensing units, one in the flow line and one in the spill line; thisincreases cost and complexity.

The use of a rate meter in series with a cumulative meter may be considered,where the accuracy of a cumulative measurement device with the ability of a ratemeter to deal with transient conditions is combined.


FlowU-shaped parallelflow tubes

Drivecoil

Right positiondetector

Magnet

Flow

Figure 12.4 Coriolis effect flow meter

Fuel consumption measurements: gaseous fuels

For these fuels, consumption measurements are made by gas flow meters. Metering ofgases is a more difficult matter than liquid metering. The density of a gas is sensitiveto both pressure and temperature, both of which must be known when, as is usuallythe case, the flow is measured by volume. Also the pressure difference availableto operate the meter is limited. The traditional domestic gas meter contains fourchambers separated by bellows and controlled by slide valves. Successive incrementsin volume metered are quite large, so that instantaneous or short-term measurementsare not possible.

Other types, capable of indicating smaller increments of flow, make use of rotorshaving sliding vanes or meshing rotors. In the case of natural gas supplied from themains, flow measurement is a fairly simple matter. Pressure and temperature at themeter must be measured and accurate data on gas properties (density, calorific value,etc.) will be available from the supplier.

Measurement of liquefied petroleum gas consumption is less straightforward, sincethis is stored as a liquid under pressure and is vaporized, reduced in pressure andheated before reaching the engine cylinders. The gas meter must be installed in theline between the ‘converter’ and the carburettor. It is essential to measure the gastemperature at this point to achieve accurate results.

The effect of fuel condition on engine performance

The effect of the condition (pressure, temperature, humidity and purity) of the com-bustion air that enters the engine is discussed later in this chapter. In the case of aspark ignition engine the maximum power output is more or less directly proportionalto the mass of oxygen contained in the combustion air, since the air/fuel ratio at fullthrottle is closely controlled.

248 Engine Testing

In the case of the diesel engine the position is different and less clear cut. Themaximum power output is generally determined by the maximum mass rate of fueloil flow delivered by the fuel injection pump operating at the maximum fuel stopposition. However, the setting of the fuel stop determines the maximum volumetricfuel flow rate and this is one reason why fuel temperature should always be closelycontrolled in engine testing.

Fuel has a high coefficient of cubical expansion, lying within the range 0.001 to0.002 per �C (compared with water, for which the figure is 0.00021 per �C). Thisimplies that the mass of fuel delivered by a pump is likely to diminish by between0.1 and 0.2 per cent for each degree rise in temperature, a by no means negligibleeffect.

A further fuel property that is affected by temperature is its viscosity, significantbecause in general the higher the viscosity of the fuel the greater the volume thatwill be delivered by the pump at a given rack setting. This effect is likely to bespecific to a particular pump design; as an example, one engine manufacturer regardsa ‘standard’ fuel as having a viscosity of 3 cSt at 40�C and applies a correctionof 2 1

2 per cent for each centistoke departure from this base viscosity, the deliveredvolume increasing with increasing viscosity.

As a further factor, the specific gravity (density) of the fuel will also affect themass delivered by a constant volume pump, though here the effect is obscured by thefact that, in general, the calorific value of a hydrocarbon fuel falls with increasingdensity, thus tending to cancel the change out.

Measurement of lubricating oil consumption

Oil consumption measurement is of increasing importance since it is an importantcomponent of the total engine emissions but it is one of the most difficult measure-ments associated with engine testing within the test cell. Oil consumption rate is veryslow relative to the quantity in circulation. In a typical vehicle engine of lubricantcapacity of 6 litres, the oil consumption rate will be in the region of zero to 25 cm3/hor up to 0.5 per cent of the volume in circulation. This means that test durations tendto be of long duration before meaningful results are obtained.

With the exception of total loss systems, such as are used for cylinder lubricationof large marine diesel engines, no entirely satisfactory method exists. One methodis to adapt the engine for dry sump operation. The sump is arranged to drain to aseparate receiver, the contents of which are monitored, either by weighing or bydepth measurement. An alternative method is to draw off oil down to a datum sumplevel and weigh it before and after the test period.

Difficulties with all systems include:

• The quantity of oil adhering to the internal surfaces of the engine is very sensitiveto temperature, as is the volume in transit to the receiver. This can give rise tolarge apparent variations in consumption rate.


• Volumetric changes due to temperature soak of the engine and fluids.• Apparent consumption is influenced by fuel dilution and by any loss of oil

vapour.• The rate of oil consumption tends to be very sensitive to conditions of load, speed

and temperature. There is also a tendency to medium-term variations in apparentrate, due to such factors as ‘ponding up’ of oil in return drains and accumulationof air or vapour in the circuit.

For a critical analysis of the dry sump method, see SAE paper 880098,1 whichalso describes a statistical test procedure aimed at minimizing these randomerrors.

An important but rather specialized technique for measuring oil consumption isby using specially blended oil containing either a sulphur compound or a radioactiveisotope. Continuous measurement of sulphur dioxide or sensing of the isotope in theexhaust allows calculation of the oil consumption rate; special sampling procedurescan be used to differentiate burned and unburned fractions.

Measurement of crankcase blow-by

In all reciprocating internal combustion engines, there is a flow of gas into andout of the clearances between piston top land, ring grooves and cylinder bore. Inan automotive gasoline engine these can amount to 3 per cent of the combustionchamber volume. Since in a spark-ignition engine the gases consist of unburnedmixture which emerges during the expansion stroke too late to be burned, this canbe a major source of HC emissions and also represents a loss of power. Some of thisgas will leak past the rings and piston skirt in the form of blow-by into the crankcase.It is then vented back into the induction manifold and to this extent reduces the HCemissions and fuel loss, but has an adverse effect on the lubricant.

Blow-by flow is highly variable over the full range of an engine’s performanceand life, therefore accurate blow-by meters will need to be able to deal with a widerange of flows and with pulsation. Instruments based on the orifice measurementprinciple, coupled with linearizing signal conditioning (the flow rate is proportionalto the square root of the differential pressure), are good at measuring the blow-by gasin both directions of flow which can occur when there is heavy pulsation betweenpressure and partial vacuum in the crankcase.

Within the gas flow there may be carbon and other ‘dirt’ particles. The sensitivityof the measuring instrument to this dirt will depend on the application; sensitivity isshown in Table 12.1.

Crankcase blow-by is a significant indicator of engine condition and should prefer-ably be monitored during any extended test sequence. An increase in blow-by canbe a symptom of various problems such as incipient ring sticking, bore polishing ordeficient cylinder bore lubrication.

250 Engine Testing

Table 12.1 Types of blow-by meters

Type Typical

accuracy

Dirt

sensitivity

Lowest flow

(l/min)

Positive impeller 2% FS Medium Approximately 6Hot wire 2% High Approximately 28Commercial gas meter 1% FS High Claimed 0.5Vortex 1% High Approximately 7Flow through an orifice 1% Low Claimed 0.2

Measurement of air consumption and gas flows

The accurate measurement of the air consumption of an internal combustion engineis a matter of some complexity but of great importance. The theory also has relevanceto gas flow measurement in emissions testing, therefore the subject is covered insome detail in the following text.

The influence of the condition of the air entering the engine on various aspectsof engine performance is discussed and ‘correction factors’ defining these effectsquantitatively are derived. The theory of various methods of measurement is givenand the limitations of each method described.

Properties of air

Air is a mixture of gases with the following approximate composition:

Major constituents By mass By volume

Oxygen, O2 23% 21%Nitrogen, N2 77% 79%

plus trace gases and water vapour of variable amount, usual range 0.2–2.0 per centof volume of dry air.

The amount of water vapour present depends on temperature and prevailingatmospheric conditions. It can have an important influence on engine performance,notably on exhaust emissions, but for all but the most precise work its influence onair flow measurement may be neglected.

The relation between the pressure, specific volume and density of air is describedby the gas equation

pa ×105 = �RTa (1)


where R, the gas constant, has the value for air

R= 287 J/kgK

A typical value for air density in temperate conditions at sea level would be

�= 1�2Kg/m3

Air consumption, condition and engine performance

The internal combustion engine is essentially an ‘air engine’ in that air is the workingfluid; the function of the fuel is merely to supply heat. There is seldom any particulartechnical difficulty in the introduction of sufficient fuel into the working cylinder,but the attainable power output is strictly limited by the charge of air that can beaspirated.

It follows that the achievement of the highest possible volumetric efficiency is animportant goal in the development of high-performance engines, and the design ofinlet and exhaust systems, valves and cylinder passages represents a major part ofthe development programme for engines of this kind.

The standard methods of taking into account the effects of charge air condition aslaid down in European and American Standards are complex and difficult to applyand are mostly used to correct the power output of engines undergoing acceptance ortype tests. A simplified treatment, adequate for most routine test purposes, is givenbelow.

The condition of the air entering the engine is a function of the followingparameters:

• pressure;• temperature;• moisture content;• impurities.

For the first three factors standard conditions, according to European and Americanpractice, are:

• atmospheric pressure 1 bar (750mmHg);• temperature 25�C (298K);• relative humidity 30%.

Atmospheric pressure

Since the volumetric efficiency of an engine tends to be largely independent of theair supply pressure, the mass of air consumed tends to vary directly with the density,

252 Engine Testing

which is itself proportional to the absolute pressure, other conditions remainingunchanged. Since the standard atmosphere = 1 bar we may write:

�t = �nP (2)

where:

�t = density under test conditions, kg/m3

�n = density under standard conditions, kg/m3

P= atmospheric pressure under test conditions, bar

It follows that a change of 1 per cent or 7.5mmHg corresponds to a change in themass of air entering the engine of 1 per cent. For most days of the year the (sealevel) atmospheric pressure will lie within the limits 750mmHg ± 3 per cent, saybetween 775 and 730mmHg, with a corresponding percentage variation in chargeair mass of 6 per cent.

It is common practice to design the test cell ventilation system to maintain a smallnegative pressure in the cell, to prevent fumes entering the control room, but the levelof depression is unlikely to exceed about 50mm water gauge (50 Pa), equivalentto a change in barometer reading of less than 0.37mmHg. Clearly the effect oncombustion air flow, if the engine is drawing air from within the cell, is negligible.

The barometer also falls by about 86mmHg (11.5 kPa) for an increase in altitudeof 1000m (the rate decreasing with altitude), see Table 12.2. This indicates that themass of combustion air falls by about 1 per cent for each 90m (300 ft) increase inaltitude, a very significant effect.

Variations in charge air pressure have an important ‘knock-on’ effect: the pres-sure in the cylinder at the start of compression will in general vary with the airsupply pressure and the pressure at the end of compression will change in the sameproportion. This can have a significant effect on the combustion process.

Table 12.2 Variation in atmospheric pressurewith height above sea level

Altitude (m) Fall in pressure (bar)

0 0500 0�059

1000 0�1151500 0�1682000 0�2183000 0�3124000 0�397


Air temperature

Variations in the temperature of the air supply have an effect of the same order ofmagnitude as variations in barometric pressure within the range to be expected intest cell operation. Air density varies inversely with its absolute temperature:

�t = �n •298

�tt +273�(3)

where tt = test temperature.However, the temperature at the start of compression determines that at the end.

In the case of a naturally aspirated diesel engine with a compression ratio of 16:1and an air supply at 25�C, the charge temperature at the start of compression wouldtypically be about 50�C. At the end of compression the temperature would be inthe region of 530�C, increasing to about 560�C for an air supply temperature 10�Chigher. The level of this temperature can have a significant effect on such factors asthe NOx content of the exhaust, which is very sensitive to peak combustion temper-ature. The same effect applies, with generally higher temperatures, to turbochargedengines.

Compared with the effects of pressure and temperature, the influence of therelative humidity of the air supply on the air charge is relatively small, except at highair temperatures. The moisture content of the combustion air does, however, exerta number of influences on performance. Some of the thermodynamic properties ofmoist air have been discussed under the heading of psychometry (Chapter 5), butcertain other aspects of the subject must now be considered.

The important point to note is that unit volume of moist air contains less oxygenin a form available for combustion than the same volume of dry air under the sameconditions of temperature and pressure. Moist air is a mixture of air and steam:while the latter contains oxygen it is in chemical combination with hydrogen and isthus not available for combustion. The European and American Standards specify arelative humidity of 30 per cent at 25�C.

This corresponds to a vapour pressure 0.01 bar, thus implying a dry air pressureof 0.99 bar and a corresponding reduction in oxygen content of 1 per cent whencompared with dry air at the same pressure. Figure 12.5 shows the variation indry air volume expressed as a percentage adjustment to the standard condition withtemperature for relative humidities ranging from 0 (dry) to 100 per cent (saturated).

The effect of humidity becomes much more pronounced at higher temperatures:thus at a temperature of 40�C the charge mass of saturated air is 7.4 per cent lessthan for dry air. The effect of temperature on moisture content (humidity) should benoted particularly, since the usual method of indicating specified moisture content fora given test method, by specifying a value of relative humidity, can be misleading.Figure 12.5 indicates that at a temperature of 0�C the difference between 0 per cent(dry) and 100 per cent (saturated) relative humidity represents a change of only 0.6per cent in charge mass; at −10�C it is less than 0.3 per cent. It follows that it is

254 Engine Testing

Change (

%)

Temperature (°C)

Rela

tive h

um

idity (

%)

+1

0

–1

–2

–3

–5

–4

–6

–7–10 0 10 20 30 40

100

80

60

40

20

0Standardcondition

Figure 12.5 Variation of air charge mass with relative humidity at differenttemperatures

barely worthwhile adjusting the humidity of combustion air at temperatures below(perhaps) 10�C.

The moisture content of the combustion air has a significant effect on the formationof NOx in the exhaust of diesel engines. The SAE procedure for measurement ofdiesel exhaust emissions gives a rather complicated expression for correcting for this.This indicates that, should the NOx measurement be made with completely dry air,a correction of the order of +15 per cent should be made to give the correspondingvalue for a test with moisture content of 60 per cent relative humidity.

Finally, it should be mentioned that pollution of the combustion air can result in areduction in the oxygen available for combustion. A likely source of such pollutionmay be the ingestion of exhaust fumes from other engines or from neighbouringindustrial processes, such as paint plant, and care should also be taken in the siting of


air intakes and exhaust discharges, to ensure that under unfavourable conditions therecannot be unintended exhaust recirculation. Exhaust gas has much the same densityas air and a free oxygen content that can be low or even zero, so that approximately1 per cent of exhaust gas in the combustion air will reduce the available oxygen inalmost the same proportion.

These various adjustments or ‘corrections’ to the charge mass are of coursecumulative in their effect and the following example illustrates the magnitudesinvolved. Consider:

A, a hot, humid summer day with the chance of thunderB, a cold, dry winter day of settled weather

Condition A B

Pressure 0.987 bar, 740mmHg 1.027 bar, 770mmHg

Temperature 35�C 10�C

Relative humidity 80% 40%

Pressure1

0�987= 1�0135

1

1�027= 0�9740

Temperature308

298= 1�0336

283

1�027= 0�9497

Relative humidity fromFig. 12.4

+3.48%= 1.0348 0%

Total adjustment 1.0840, +8.4% 0.9233, −7.7%

This ‘adjustment’ indicates the factor by which the observed power should bemultiplied to indicate the power to be expected under ‘standard’ conditions. We seethat under hot, humid, stormy conditions, the power may be reduced by as much as15 per cent compared with the power under cold anticyclonic conditions, a by nomeans negligible adjustment.

It should be pointed out that the power adjustment calculated above is basedon the assumption that charge air mass directly determines the power output, butthis would only be the case if the air/fuel ratio were rigidly controlled (as in sparkignition engines with precise stoichiometric control). In most cases a reduction, forexample, in charge air mass due to an increase in altitude will result in a reductionin the air/fuel ratio, perhaps with an increase in exhaust smoke, but not necessarilyin a reduction in power. The correction factors laid down in the various standardstake into account the differing responses of the various types of engine to changesin charge oxygen content.

It should also perhaps be pointed out that the effects of water injection are quitedifferent from the effects of humidity already present in the air. Humid air is amixture of air and steam: the latent heat required to produce the steam has alreadybeen supplied. When, as is usually the case, water is injected into air leaving the

256 Engine Testing

turbocharger at a comparatively high temperature, the cooling effect associated withthe evaporation of the water achieves an increase in charge density which muchoutweighs the decrease associated with the resulting steam.

The airbox method of measuring air consumption

The simpler methods of measuring air consumption involve drawing the air throughsome form of measuring orifice and measuring the pressure drop across the orifice.2

It is good practice to limit this drop to not more than about 125mm H2O (1200 Pa).For pressures less than this, air may be treated as an incompressible fluid, with muchsimplification of the air flow calculation.

The velocity U developed by a gas expanding freely under the influence of apressure difference �p, if this difference is limited as above, is given by:

�U 2

2= �p�U =

√

2�p

�(4)

Typically, air flow is measured by means of a sharp-edged orifice mounted in theside of an airbox, coupled to the engine inlet and of sufficient capacity to damp outthe inevitable pulsations in the flow into the engine, which are at their most severe inthe case of a single cylinder four-stroke engine, Fig. 12.6. In the case of turbochargedengines, the inlet air flow is comparatively smooth and a well-shaped nozzle withoutan airbox will give satisfactory results.

The air flow through a sharp-edged orifice takes the form sketched in Fig. 12.7.The coefficient of discharge of the orifice is the ratio of the transverse area of theflow at plane a (the vena contracta) to the plan area of the orifice. Tabulated valuesof Cd are available in BS 1042,3 but for many purposes a value Cd = 0.60 may beassumed.

We may easily derive the volumetric flow rate of air through a sharp-edged orificeas follows:

Flow rate= coefficient of discharge× cross-sectional area of orifice× velocityof flow

Q= Cd

�d2

4

√

2�p

�(5)

Noting from eq. (3) that

�=pa ×105

RTa


Air flo

w r

ate

(nom

inal scale

)

Crank angle (degrees)

3.000

2.500

2.000

1.500

1.000

0.500

0.000

–0.500

–1.000b.d.c.

0t.d.c.180

t.d.c.540

b.d.c.720

Meanflow rate

Zero flow

b.d.c.360

Figure 12.6 Induction air flow, single cylinder, four-stroke diesel engine

Venacontractaa

a

Figure 12.7 Flow through a sharp-edged orifice

258 Engine Testing

and assuming that �p is equivalent to hmm H2O, we may write

Q= Cd

�d2

4

√

2×9�81h×287T

pa ×105(6a)

Q= 0�1864Cdd2

√

hTa

pa

m3/s (6b)

To calculate the mass rate of flow, note that:

m= �Q=paQ

RTa

giving, from eq. (6a):

m= Cd

�d2

4

√

2×9�81h×105

287Ta

(7a)

m= 64�94Cdd2

√

hpa

Ta

(7b)

Equations (6b) and (7b) give the fundamental relationship for measuring air flow byan orifice, nozzle or venturi.

Sample calculation

If

Cd = 0�6;d = 0�050m;h= 100mmH2O;Ta = 293K (20�C);pa = 1�00 bar.

Then

Q= 0�04786m3/s �1�69 ft3/s�m= 0�05691kg/s �0�1255 lb/s�

To assist in the selection of orifice sizes, Table 12.3 gives approximate flow ratesfor orifices under the following standard conditions:


Table 12.3 Approximate air flow ratesfor orifices sizes 10 to 150mm

Orifice dia. (mm) Q (m3/s) m (kg/s)

10 0�002 0�00220 0�008 0�00950 0�048 0�057

100 0�19 0�23150 0�43 0�51

h = 100 mmH2OTa = 293 K (20�C)pa = 1.00 bar

A disadvantage of flow measurement devices of this type is that the pressuredifference across the device varies with the square of the flow rate. It follows thata turndown in flow rate of 10:1 corresponds to a reduction in pressure difference of100:1, implying insufficient precision at low flow rates. It is good practice, when awide range of flow rates is to be measured, to select a range of orifice sizes, eachcovering a turndown of not more than 2.5:1 in flow rate.

Air consumption of engines approximate calculation

The air consumption of an engine may be calculated from:

V = v

Vs

K

n

60m3/s (8)

where K= 1 for a two-stroke and 2 for a four-stroke engine.For initial sizing of the measuring orifice v, the ratio of the volume of air

aspirated per stroke to the volume of the cylinder may be assumed to be about 0.8for a naturally aspirated engine and up to about 2.5 for supercharged engines.

Sample calculation

Single cylinder four-stroke engine, swept volume 0.8 litre, running at a maximum of3000 rev/min, naturally aspirated:

V = 0�80�0008

2

3000

60= 0�16m3/s

Suitable orifice size, from Table 12.3, 30mm.

260 Engine Testing

Connection of airbox to engine inlet

It is essential that the configuration of the connection between the airbox and theengine inlet should model as closely as possible the configuration of the air intakearrangements in service. This is because pressure pulsations in the inlet can have apowerful influence on engine performance, in terms of both volumetric efficiencyand pumping losses.

The resonant frequency of a pipe of length L, open at one end and closed atthe other, = a/4L, where ‘a’ is the speed of sound, roughly 330m/s. Thus an inletconnection 1m long would have a resonant frequency of about 80Hz. This corre-sponds to the frequency of intake valve opening in a four cylinder four-stroke enginerunning at 2400 rev/min. Clearly such an intake connection could disturb engineperformance at this speed. In general, the intake connection should be as short aspossible.

The viscous flow air meter

The viscous flow air meter, invented by Alcock and Ricardo in 1936, was for manyyears the most widely used alternative to the airbox and orifice method of measuringair flow. In this device the measuring orifice is replaced by an element consisting of alarge number of small passages, generally of triangular form. The flow through thesepassages is substantially laminar, with the consequence that the pressure differenceacross the element is approximately proportional to the velocity of flow, rather thanto its square, as is the case with a measuring orifice.

This has two advantages. Firstly average flow is proportional to average pressuredifference, implying that a measurement of average pressure permits a direct calcu-lation of flow rate, without the necessity for smoothing arrangements. Secondly, theacceptable turndown ratio is much greater.

The flow meter must be calibrated against a standard device, such as a measuringorifice.

Lucas–Dawe air mass flow meter

This device depends for its operation on the corona discharge from an electrodecoincident with the axis of the duct through which the air is flowing. Air flowdeflects the passage of the ion current to two annular electrodes and gives rise to animbalance in the current flow that is proportional to air flow rate.

An advantage of the Lucas–Dawe flow meter is its rapid response to changes inflow rate, of the order of 1 ms, making it well suited to transient flow measurements,but it is sensitive to air temperature and humidity, and requires calibration against astandard.


Positive displacement flow meters

As rotation occurs, successive pockets of air are transferred from the suction to thedelivery side of the flow meter, and flow rate is proportional to rotor speed. Someof these flow meters operate on the principle of the Roots blower. Advantages ofthe positive displacement flow meter are accuracy, simplicity and good turndownratio. Disadvantages are cost, bulk, relatively large pressure drop and sensitivity tocontamination in the flow.

Hot wire or hot film anemometer devices

The principle of operation of these popular devices is based on the cooling effectcaused by gas flow on a heated wire or film surface. The heat loss is directlyproportional to the air mass flow rate, providing the flow through the device islaminar. The advantages of these designs are their reliability and good tolerance tocontaminated air flows. However, calibration at site is virtually impossible so theyare supplied with maker’s certification. Probably the best known device of this typein European test cells is the Sensyflow™ range made by ABB.

Notation

Atmospheric pressure pa barunder test conditions p bar

Atmospheric temperature Ta KTest temperature t�tCDensity of air � kg/m3

under standard conditions �n kg/m3

under test conditions �t kg/m3

Pressure difference across orifice �� Pa, h mmH2OVelocity of air at contraction U m/sCoefficient of discharge of orifice Cd

Diameter of orifice d mVolumetric rate of flow of air Q m3/sMass rate of flow of air m kg/sConstant, 1 for two-stroke and 2 for four-stroke engines K

Engine speed n rev/minNumber of cylinders Nc

Swept volume, total Vs m3

Air consumption rate V m3/sVolumetric efficiency of engine v

Volume of airbox Vb m3

Gas constant R J/kg K

262 Engine Testing

References

1. Johren, P.-W. and Newman, B.A. (1988) Evaluating the oil consumption behaviour ofreciprocating engines in transient operation, SAE Paper 880098.

2. Kastner, L.J. (1947) The airbox method of measuring air consumption, Proc. I. Mech.

E., 157.3. BS 1042 Measurement of Fluid Flow in Closed Conduits: Section 1.1, (AS 2360.1.2-

1993) Specification for Square-edged Orifice Plates, Nozzles and Venturi Tubes Inserted in

Circular Cross-section; Conduits Running Full; Section 1.4, Guide to the Use of Devices

Specified in Sections 1.1 and 1.2.

Further reading

BS 5514 Parts 1 to 6 Reciprocating Internal Combustion Engines: Performance.BS 7405 Guide to Selection and Application of Flowmeters for the Measurement of

Fluid Flow in Closed Conduits.Plint, M.A. and Böswirth, L. (1978) Fluid Mechanics: A Laboratory Course, Griffin,

London.

13 Thermal efficiency, measurementof heat and mechanical losses

Introduction

This chapter deals with the measurements and calculations necessary to determine theenergy balance, thermal performance and mechanical losses of an internal combustionengine. A brief account is given of the basic theory in order to provide a frameworkfor an interpretation of these observations and as background reading for Chapter 14,where modern combustion analysis is discussed. A notation table is included at theend of the chapter.

One ultimate measure of the performance of an internal combustion engine is theproportion of the heat of combustion of the fuel that is turned into useful work at theengine coupling. The thermal efficiency at full load of internal combustion enginesranges from about 20 per cent for small gasoline engines up to more than 50 per centfor large slow-running diesel engines, which are the most efficient means currentlyavailable of turning the heat of combustion of fuel into mechanical power.

It is useful to have some idea of the theoretical maximum thermal efficiencythat is possible, as this sets a target for the engine developer. Theoretical thermo-dynamics allows us, within certain limitations, to predict this maximum value. Theproportion of the heat of combustion that is not converted into useful work appearselsewhere: in the exhaust gases, in the cooling medium and as convection and radi-ation from the hot surfaces of the engine. In addition, there may be appreciablelosses in the form of unburned or late burning fuel. It is important to be able toevaluate these various losses. Of particular interest are losses from the hot gas in thecylinder to the containing surfaces, since these directly affect the indicated powerof the engine. The so-called ‘adiabatic engine’ seeks to minimize these particularlosses.

Some of the heat carried away in the exhaust gas may be converted into usefulwork in a turbine or used for such purposes as steam generation or the productionof hot water.

264 Engine Testing

Fundamentals

Calorific value of fuels

The calorific value of a fuel is defined in terms of the amount of heat liberated whena fuel is burned completely in a calorimeter. Detailed methods and definitions aregiven in Ref. 1, but for the present purpose the following is sufficient.

Since all hydrocarbon fuels produce water as a product of combustion, part ofthese products (the exhaust gas in the case of an i.c. engine) consists of steam. If,as is the case in a calorimeter, the products of combustion are cooled to ambienttemperature, this steam condenses, and in doing so gives up its latent heat. Thecorresponding measure of heat liberated is known as the higher or gross calorificvalue (also known as gross specific energy). If no account is taken of this latent heatwe have the lower or net calorific value (also known as net specific energy). Sincethere is no possibility of an internal combustion engine making use of the latentheat, it is the invariable practice to define performance in terms of the lower calorificvalue C1. Table 13.1 shows values of the lower calorific value and density for sometypical fuels.

Gaseous fuels

These fuels, which have favourable emissions characteristics, are becoming ofincreasing importance:2�3

• Natural gas (NG), also sometimes described as compressed natural gas (CNG)and, when transported in bulk at very low temperature, as liquefied natural gas(LNG);

Table 13.1 Properties of liquid fuels

Lower calorific

value∗ (MJ/kg)

Stoichiometric

air/fuel ratio

Density

(kg/l)

Gasoline 43�8 14�6 0�74Gas oil 42�5 14�8 0�84Methanol 19�9 6�46 0�729Ethanol 27�2 8�94 0�79Light fuel oil 40�6 0�925Medium fuel oil 39�9 14�4 0�95Heavy fuel oil 39�7 0�965

∗ BS 5514 Part 1 Reciprocating Internal Combustion Engines: Performancespecifies a standard lower calorific value for distillate fuels as 42.7MJ/kg.

Thermal efficiency, measurement of heat and mechanical losses 265

Table 13.2 Approximate properties of typical gaseous fuels3

Natural gas (North Sea gas)

Methane CH4 93.3%Higher hydrocarbons 4.6%N2, CO2 2.1%Lower calorific value∗ 48.0MJ/kgStoichiometric air/fuel ratio 14.5 : 1Approximate density (gas at 0�C) 0.79 kg/m3

Liquefied petroleum gas (LPG)

Propane C3H8 90%Butane C4H10 5%Unsaturates 5%Lower calorific value 46.3MJ/kgStoichiometric air/fuel ratio 15.7 : 1Approximate density (gas at 0�C) 2.0 kg/m3

∗ It should be noted that some gas suppliers commonly quote higher as opposed to lowercalorific value. In the case of methane, with its high H/C ratio, the difference is nearly 10%.

• natural gas consists mainly of methane but, having evolved from organic deposits,invariably contains some higher hydrocarbons and traces of N2 and CO2.Composition varies considerably from field to field and the reader involved inwork on natural gas should establish the particulars of the gas with which he isconcerned.

• North Sea gas has become a standard gaseous fuel in the UK and its approximateproperties are shown in Table 13.2.

Liquefied petroleum gas (LPG or LP-gas)

LPG is a product of the distillation process of crude oil or a condensate from wetnatural gas. It consists largely of propane and, unlike natural gas, can be stored inliquid form at moderate pressures. In view of its good environmental properties (lowunburned hydrocarbon emissions, low CO, virtually no particulate emissions and nosulphur), it is finding favour as a vehicle engine fuel. NOx emissions tend to behigher than for gasoline, owing to its high combustion temperature. It has a highoctane number, RON 110, which permits higher compression ratios. Approximateproperties are shown in Table 13.2.

Ideal standard cycles: effect of compression ratio

Many theoretical cycles for the internal combustion engine have been proposed,some of them taking into account such factors as the exact course of the combustion

266 Engine Testing

process, the variation of the specific heat of air with temperature and the effects ofdissociation of the products of combustion at high temperature. However, all thesecycles merely modify the predictions of the cycle we shall be considering, generallyin the direction of reduced attainable efficiency, without much changing the generalpicture; for a detailed discussion, see Heywood.4

The air standard cycle, also known as the Otto cycle, is shown in Fig. 13.1. Itconsists of four processes, forming a complete cycle, and is based on the followingassumptions:

1. The working fluid throughout the cycle is air, and this is treated as a perfect gas.2. The compression process 1–2 and the expansion process 3–4 are both treated as

frictionless and adiabatic (without heat loss).3. In place of heat addition by internal combustion phase 2–3 of the process is

represented by the addition of heat from an external source, the volume remainingconstant.

4. The exhaust process is replaced by cooling at constant volume to the initialtemperature.

It will be evident that these conditions differ considerably from those encoun-tered in an engine; nevertheless the thermodynamic analysis, which is very simple,gives useful indications regarding the performance to be expected from an inter-nal combustion engine, in particular with regard to the influence of compressionratio.

The air standard cycle efficiency is given by:

�as = 1−1

R�−1(1a)

p

3

2 4

1

V

Figure 13.1 Air standard cycle


The course of events in an engine cylinder departs from this theoretical pattern inthe following main respects:

1. Heat is lost to the cylinder walls, reducing the work necessary to compressthe air, the rise in temperature and pressure during combustion, and the workperformed during expansion.

2. Combustion, particularly in the diesel engine, does not take place at constantvolume, resulting in a rounding of the top of the diagram, point 3, and areduction in power. A better standard of reference for the diesel engine isthe limited pressure cycle, Fig. 13.2, for which the efficiency is given by theexpression:

�= 1−1

R�−1

[

�� −1

�� −1+�−1

]

(1b)

where

�=p3

p2a�=

V3bV3a

this reduces to eq. (1a), when �= 1.3. The properties of air and of the products of combustion, do not correspond to

those of an ideal gas, resulting in a smaller power output than predicted.4. The gas exchange process is ignored in the standard cycle.

Figure 13.3 shows the variation of air standard cycle efficiency with compressionratio and shows the range of this ratio for spark ignition and diesel engines. It isclearly desirable to use as high a compression ratio as possible. Also shown is the

p3a 3b

4

2

1

V

Figure 13.2 Limited pressure cycle

268 Engine TestingT

herm

al effic

iency

1.0

0.8

0.6

0.4

0.2

0

Compression ratioOtto Diesel

0 2 4 6 8 10 12 14 16 18 20 22

2–5 l capacity

Real engines

Air standard cycle

Figure 13.3 Variation of air standard cycle efficiency with compression ratio

approximate indicated thermal efficiency to be expected from gasoline and dieselengines of 2–5-litre swept volume. Larger engines and in particular large slow-speed diesel engines can achieve significantly higher efficiencies, mainly becauseheat losses from the cylinder contents become less in proportion as the size of theindividual cylinders increases.

The energy balance of an internal combustion engine

The distribution of energy in an internal combustion engine is best considered interms of the steady flow energy equation, combined with the concept of the controlvolume. In Fig. 13.4 an engine is shown, surrounded by the control surface. Thevarious flows of energy into and out of the control volume are shown.

In

• fuel, with its associated heat of combustion;• air, consumed by the engine.

Out

• power developed by the engine;• exhaust gas;• heat to cooling water or air;• convection and radiation to the surroundings.


Convectionand radiation

Fuel

Air

In

Coolingwater

Out

Oil cooler

Exhaust

Power

Cooling water

Engine

In Out

Figure 13.4 Control volume of an i.c. engine showing energy flows

The steady flow energy equation gives the relationship between these quantities, andis usually expressed in kilowatts:

H1 = Ps+ �H2−H3+Q1+Q2 (2)

in which the various terms have the following meanings:

H1 = combustion energy of fuel = mfCL×103

Ps = power output of engineH2 = enthalpy of exhaust gas∗ = �mf + maCpTe

H3 = enthalpy of inlet air = maCpTa

Q1 = heat to cooling water† = mWCW �T2W−T1W

Q2 = convection and radiation.

This assumes that the specific heat of the exhaust gas, the mass of which is the sumof the masses of air and fuel supplied to the engine, is equal to that of air. This is notstrictly true, but permits an approximate calculation to be made if the temperature ofthe exhaust gas is measured (exact measurement of exhaust temperature is no simplematter, see Chapter 20).

Note that it is not possible to show the indicated power directly in this energybalance since the difference between it and the power output Ps representing friction

∗ For detailed description of enthalpy see, for example, Refs 4 and 5.† This may also include heat transferred to the lubricating oil and subsequently to the coolingwater via an oil cooler.

270 Engine Testing

and other losses appears elsewhere as part of the heat to the cooling water Q1 andother losses Q2.

Measurement of heat losses: heat to exhaust

If air and fuel flow rates, also exhaust temperature, are known, this may be calculatedapproximately, see H2 above. For an accurate measurement of exhaust heat, use canbe made of an exhaust calorimeter. This is a gas-to-water heat exchanger in whichthe exhaust gas is cooled to a moderate temperature and the heat content measuredfrom observation of cooling water flow rate and temperature rise.

The expression for H2 becomes:

H2 = mcCW �T2c−T1c+ �mf + maCpTco (3)

The rate of flow of cooling water through the calorimeter should be regulated so thatthe temperature of the gas leaving the calorimeter, Tco, does not fall below about60�C (333K). This is approximately the dew point temperature for exhaust gas: atlower temperatures the steam in the exhaust will start to condense, giving up itslatent heat, see section on Calorific value of fuels.

Sample calculation: analysis of an engine test

Table 13.3 is an analysis, based on eq. (2), of one test point in a sequence of testson a vehicle engine.

Table 13.3 Energy balance of a gasoline engine at full throttle(four cylinder, four-stroke engine, swept volume 1.7 litre)

Engine speed 3125 rev/minPower output Ps = 36.8 kWFuel consumption rate mf = 0.00287 kg/sAir consumption rate ma = 0.04176 kg/sLower calorific value of fuel CL = 41.87× 106 J/kgExhaust temperature Tc = 1066K (793�C)Cooling water flow mw = 0.123 kg/sCooling water inlet temperature∗ T1w = 9.2�CCooling water outlet temperature∗ T2w = 72.8�CInlet air temperature Ta = 292 K (19�C)

∗ The engine was fitted with a heat exchanger. These are the temperaturesof the primary cooling water flow to the exchanger.


Then noting that:

Specific heat of air at constant pressure Cp = 100 kJ/kg KSpecific heat of water Cw= 4.18 kJ/kg KH1 = 0.00287× 41.87× 103 = 120.2 kWPs = 36.8 kWH2 = (0.00287+ 0.04176)× 1.00× 1066 = 47.6 kWH3 = 0.04176× 1.00× 292 = 12.2 kWH2 – H3 = 35.4 kWQ1 = 0.123× 4.18 (72.8− 9.2) = 32.7 kWQ2 (by difference) = 15.3 kW

We may now draw up an energy balance (quantities in kilowatts):

Heat of combustion H1 120.2 Power output Ps 36�8 �30�6%

Exhaust (H2−H3) 35�4 �29�5%

Other losses Q 15�3 �12�7%

120.2 120�2 �100%

The thermal efficiency of the engine

�th =Ps

H1

= 0�306

The compression ratio of the engine R= 8.5, giving:

�as = 1−1

8�51�4−1= 0�575

The mechanical efficiency of this engine at full throttle was approximately 0.80,giving an indicated thermal efficiency of

0�306

0�8= 0�3825

This is approximately two thirds of the air standard efficiency.

Sample calculation: exhaust calorimeter

In the test analysed in Table 13.3, the heat content of the exhaust was also measuredby an exhaust calorimeter with the following result:

cooling water flow, mc = 0.139 kg/scooling water inlet temperature, T1c = 9.2�Ccooling water outlet temperature, T2c = 63.4�Cexhaust temperature leaving calorimeter, Tco = 355K (82�C)

272 Engine Testing

Then from eq. (3):

H2 = 0�139×4�18 �63�4−9�2+ �0�00287+0�04176×1�00×355= 47�3kW

H3 = 12�2 kW, as before, giving heat to exhaust H2−H3 = 35�1 kW. This shows asatisfactory agreement with the approximate value derived from the exhaust temper-ature and air and fuel flow rates.

Energy balances: typical values

Figure 13.5 shows full power energy balances for several typical engines. Resultsfor the gasoline engine analysed above are shown at (a).

Table 13.4 shows energy balances in terms of heat losses per unit power outputfor various engine types. This will be found useful when designing such test cellservices as cooling water and ventilation. They are expressed in terms of kW/kWpower output.

1.00

0.80

0.60

0.40

0.20

0.00(a)

Pro

port

ion o

f heat re

leased b

y fuel

(b) (c) (d)

Power output

Heat to charge air

Heat to cooling water

Heat to exhaust

Heat to oil cooler

Convection and radiation

Figure 13.5 Typical full power energy balances: (a) 1.71 gasoline engine(1998); (b) 2.51 naturally aspirated diesel engine; (c) 200 kW medium speedturbocharged marine diesel; (d) 7.6MW combined heat and power unit


Table 13.4 Energy balance (kW per kW power output)

Automotive

gasoline

Automotive

diesel

Medium speed

heavy diesel

Power output 1�0 1�0 1�0Heat to cooling water 0�9 0�7 0�4Heat to oil cooler 0�05Heat to exhaust 0�9 0�7 0�65Convection and radiation 0�2 0�2 0�15

Total 3�0 2�6 2�2

These proportions depend on the thermal efficiency of the engine and are only anapproximate guide.

The role of indicated power in the energy balance

It should be observed that the indicated power output of the engine does not appearin any of our formulations of the energy balance. There is a good reason for this:the difference between indicated and brake power represents the friction losses inthe engine and the power required to drive the auxiliaries, and it is impossible toallocate these between the various heat losses in the balance.

Most of the friction losses between piston and cylinder will appear in the coolingwater; bearing losses and the power required to drive the oil pump will appear mostlyin the oil cooler, while water pump losses will appear directly in the cooling water.An exact analysis is problematic and beyond the scope of this book.

Sample calculation: prediction of energy balance

Chapter 2 deals with the concept of the engine test cell as a thermodynamic systemand the recommendation is made that at an early stage in designing a new test cellan estimate should be made of the various flows: fuel, air, water, electricity, heatand energy into and out of the cell.

In such cases it is usually necessary to make some assumptions as to the fullpower performance of the largest engine to be tested. This is possible on the basisof information given earlier and in this chapter and an example follows.

Prediction of energy balance

Taking the example of a 250 kW turbocharged diesel engine at full power. Assumespecific fuel consumption = 0.21 kg/kWh (LCV 40.6MJ/kg; thermal efficiency 0.42).Then following the general recommendations of Table 13.4, we may make an esti-mate, summarized in Table 13.5.

274 Engine Testing

The corresponding thermodynamic system is shown in Fig. 13.6. This also showsrates of flow of fuel, air and cooling water, based on the following estimates:

Fuel flow

Assume fuel density 0.9 kg/litre

Fuel flow=250×0�21

0�9= 581 l/hour �52�5kg/hour

Table 13.5 Energy balance, 250 kW turbocharged diesel engine

In Out

Fuel 592 kW Power 250 kW (42.2%)Heat to cooling water 110 kW (18.6%)Heat to oil cooler 15 kW (2.5%)Heat to exhaust 177 kW (29.9%)Convection and radiation 40 kW (6.8%)

592 kW 592 kW

Convection and radiation40 kW

Fuel52 kg/h592 kW

Air1312 kg/h

Engine

Exhaust1365 kg/h177 kW

Power250 kW

Oil cooler

Coolingwater

15 kW1300 kg/h

Coolingwater9500 kg/h

110 kW

Figure 13.6 Control volume, 250 kW diesel engine showing energy and fluidflows


Induction air flow

Assume full load air/fuel ratio = 25 : 1Air flow= 250× 0.21× 25= 1312.5 kg/hTaking air density as 1.2 kg/m3

Air flow= 1094m3/h (0.30 m3/s, 10.7 ft3/s)

Cooling water flow

Assume a temperature rise of 10�C through the jacket and oil cooler. Then sincethe specific heat of water 4.18 kJ/kg �C and 1 kWh= 3600 kJ:flow to jacket + oil cooler

125×60

4�18×10= 180kg/min �180 l/min

Exhaust flow

Sum of fuel flow + induction air flow= 1312.5+ 52.5= 1365 kg/h.

Energy balance for turbocharger6

In the case of turbocharged engines, it is useful to separate the energy flows to andfrom the turbocharger and associated air cooler from those associated with the com-plete engine. In this way an energy balance may be drawn up covering the following:

• exhaust gas entering the turbine from engine cylinders;• induction air entering the compressor;• cooling water entering the air cooler;• exhaust leaving the turbine;• induction air leaving the cooler;• cooling water leaving the cooler.

A separate control volume contained within the control volume for the completeengine may be defined and an energy balance drawn up.

Summary

Stages in calculation of energy balance from an engine test

1. Obtain information on the lower calorific value of the fuel.2. From a knowledge of compression ratio of engine calculate air standard cycle

efficiency as a yardstick of performance.3. For one or a number of test points measure:

• fuel consumption rate;• air flow rate to engine;

276 Engine Testing

• exhaust temperature;• power output;• cooling water inlet and outlet temperatures and flow rate.

4. Calculate the various terms of the energy balance and the thermal efficiencyfrom this data.

Prediction of energy balance for a given engine

1. Record type of engine and rated power output.2. Calculate fuel flow rate from known or assumed specific fuel consumption.3. Draw up energy balance using the guidelines given in Table 13.4.

Measurement of mechanical losses in engines

This section is devoted to a critical study of the various methods of determining themechanical efficiency of an internal combustion engine.

It is a curious fact that, in the long run, all the power developed by all the roadvehicle engines in the world is dissipated as friction: either mechanical friction inthe engine and transmission, rolling resistance between vehicle and road or windresistance.

Mechanical efficiency, a measure of friction losses in the engine, is thus animportant topic in engine development and therefore engine testing. It may exceed80 per cent at high power outputs, but is generally lower and is of course zero whenthe engine is idling.

Under mixed driving conditions for a passenger vehicle between one third and onehalf of the power developed in the cylinders is dissipated either as mechanical frictionin the engine, in driving the auxiliaries such as alternator and fan, or as pumping lossesin the induction and exhaust tracts. Since the improvement of mechanical efficiencyis such an important goal to engine and lubricant manufacturers, an accurate measureof mechanical losses is of prime importance. In fact the precise measurement ofthese losses is a particularly difficult problem, to which no completely satisfactorysolution exists.

Fundamentals

The starting point in any investigation of mechanical losses should ideally be a preciseknowledge of the power developed in the engine cylinder. This is represented by theindicator diagram, Fig. 13.7, which shows the relation between the pressure of thegas in the cylinder and the piston stroke or swept volume (also see Chapter 14). Fora four-stroke engine, account must be taken of both the positive area A1, representingthe work done on the piston during the compression and expansion strokes, and


p

p0

Vc Vs

A2

A1

V

1

4

3

2

Figure 13.7 Indicator diagram, four-stroke engine. 1, induction; 2, compres-sion; 3, expansion; 4, exhaust

the negative area A2, representing the ‘pumping losses’, the work performed bythe piston in expelling the exhaust gases and drawing the fresh charge into thecylinder.

In the case of an engine fitted with a mechanical supercharger or an exhaust gasturbocharger, there are additional exchanges of energy between the exhaust gas andthe fresh charge, but this does not invalidate the indicator diagram as a measure ofpower developed in the cylinder.

The universally accepted measure of indicated power is the indicated mean effec-tive pressure (i.m.e.p.). This represents the mean positive pressure exerted on thepiston during the working strokes after allowing for the negative pressure representedby the pumping losses.

The relation between i.m.e.p. and indicated power of the engine is given by theexpression:

Pi =piVsn

60K×10−1 kW (4)

here n/60K represents the number of power strokes per second in each cylinder.The useful power output of the engine may be represented by the brake mean

effective pressure

P =pbVsn

60K×10−1 kW (5)

278 Engine Testing

The mechanical losses in the engine plus power to drive auxiliaries are representedby (Pi – P) and may be represented by the friction mean effective pressure

pf = pi− pb

It may not be realistic, for several experimental reasons, to place much faith inmeasurements of indicated power based on the indicator diagram. Several other tech-niques, each having their limitations, will be discussed and the critical determinationof top dead centre (t.d.c.) is covered in detail in Chapter 14.

Motoring tests

One method of estimating mechanical losses involves running the engine under stabletemperature conditions and connected to a four-quadrant dynamometer. Ignition orfuel injection are then cut and the quickest possible measurement made of the powernecessary to motor the engine at the same speed.

Sources of error include:

• Under non-firing conditions, the cylinder pressure is greatly reduced, with aconsequent reduction in friction losses between piston rings, cylinder skirt andcylinder liner and in the running gear.

• The cylinder wall temperature falls very rapidly as soon as combustion ceases,with a consequent increase in viscous drag that may to some extent compensatefor the above effect.

• Pumping losses are generally much changed in the absence of combustion.

Many detailed studies of engine friction under motored conditions have been reportedin the literature, for a summary see Ref. 7. These usually involved the progressiveremoval of various components: camshaft and valve train, oil and fuel pumps, waterpump, generator, seals, etc., in order to determine the contribution made by eachelement.

The Morse test

In this test, the engine is run under steady conditions and ignition or injection iscut off in each cylinder in turn: it is of course only applicable to multicylinderengines.

On cutting out a cylinder, the dynamometer is rapidly adjusted to restore theengine speed and the reduction in power measured. This is assumed to be equal tothe indicated power contributed by the non-firing cylinder. The process is repeatedfor all cylinders and the sum of the reductions in power is taken to be a measure ofthe indicated power of the engine.


A modification of the Morse test8 makes use of electronically controlled unitinjectors, allowing the cylinders to be disabled in different ways and at differentfrequencies, thus keeping temperatures and operating conditions as near normal aspossible.

The Morse test is subject, though to a less extent, to the sources of error describedfor the motoring test.

The Willan’s line method

This is applicable only to unthrottled compression ignition engines. It is a matterof observation that a curve of fuel consumption rate against torque or b.m.e.p. atconstant speed plots quite accurately as a straight line up to about 75 per cent of fullpower, Fig. 13.8. This suggests that for the straight line part of the characteristic,equal increments of fuel produce equal increments of power; combustion efficiencyis constant.

At zero power output from the engine, all the fuel burned is expended in over-coming the mechanical losses in the engine, and it is a reasonable inference that anextrapolation of the Willan’s line to zero fuel consumption gives a measure of thefriction losses in the engine.

Strictly speaking, the method only allows an estimate to be made of mechanicallosses under no-load conditions. When developing power the losses in the enginewill undoubtedly be greater.

b.m.e.p. (bar)

–3 –2 –1 0 1 2 3 4 5 6 7

f.m.e.p.

3

2

1

Fuel consum

ption (

l/h)

Figure 13.8 Willan’s line for a diesel engine

280 Engine Testing

Summary

The four standard methods of estimating mechanical losses in an engine and itsauxiliaries have been briefly described. No great accuracy can be claimed for any ofthese methods and it is instructive to apply as many of them as possible and comparethe results (for a critical assessment, see Ref. 9). Measurement of mechanical lossesin an engine is still something of an ‘art’.

While no method can be claimed to give a precise absolute value for mechanicallosses they are, of course, quite effective in monitoring the influence of specificchanges made to a particular engine.

The intention is to give non-expert readers the essential background to this area ofengine performance testing so that the use of modern tools used in engine indicatingand combustion analysis described briefly in Chapter 14 can be better appreciated.

Notation

Indicated mean effected pressure (i.m.e.p.) pi barBrake mean effected pressure (b.m.e.p.) pb barFriction mean effected pressure (f.m.e.p.) pf barSwept volume of engine Vs litreEngine speed n rev/minConstant: 1 for two-stroke, 2 for four-stroke engines K

Indicated power Pi kWEngine power output P kW

Thermal efficiency

Lower calorific value CL MJ/kgCompression ratio R

Ratio of specific heat of air � �=1�4Air standard efficiency �as

Mass flow rate of fuel mf kg/sMass flow rate of inlet air ma kg/sPower output of engine Ps kWSpecific heat of air at constant pressure Cp kJ/kg KSpecific heat of water Cw kJ/kg KAmbient temperature Ta KExhaust temperature Te KCooling water inlet temperature T1w KCooling water outlet temperature T2w KMass flow rate of cooling water mW kg/s


For exhaust calorimeter

Mass flow rate of cooling water mC kg/sCooling water inlet temperature T1C KCooling water outlet temperature T2C KTemperature of exhaust leaving calorimeter TCO K

References

1. BS 7420 Guide for Determination of Calorific Values of Solid, Liquid and Gaseous Fuels

(including Definitions).2. A.S.T.M. Special Technical Publication 525 (1972) LP-Gas Engine Fuels, American Society

for Testing and Materials.3. Goodyear, E.M. (1980) Alternative Fuels, Cranfield, Macmillan, London.4. Heywood, J.B. (1988) Internal Combustion Engine Fundamentals, McGraw-Hill, Maiden-

head.5. Eastop, T.D. and McConkey, A. (1993) Applied Thermodynamics for Engineering Tech-

nologists, Longman, London.6. Watson, N. and Janota, M.S. (1982) Turbocharging the Internal Combustion Engine, Wiley-

Interscience, New York.7. Martin, F.A. (1985) Friction in Internal Combustion Engine Bearings, I. Mech. E. Paper

C 67/85.8. Haines, S.N.M. and Shields, S.A. (1989) The determination of diesel engine friction char-

acteristics by electronic cylinder disablement, Proc. I. Mech. E. Part A, 203 (A2), 129–138.9. Ciulli, E. (1993) A review of internal combustion engine losses. Part 2: Studies of global

evaluations, Proc. I. Mech. E. Part D, 207 (D3), 229–240.

14 The combustion process andcombustion analysis

Introduction

This chapter deals with the process of combustion in the internal combustion engineand the equipment available to the test engineer for its analysis. The processes takingplace in the cylinder and combustion chamber are central to the performance of theinternal combustion engine; in fact, the rest of the engine may be regarded merelyas a device for managing these processes and for extracting useful work from them.

The term ‘combustion analysis’ (CA) has become somewhat interchangeable withthat of ‘engine indicating’ (EI), the latter term covers both combustion analysis and anumber of other engine phenomena that occur at the same timescale as combustion,such as injector needle movement. In this book the term ‘engine indicating’ is usedto cover the general task of high speed engine data analysis and the term ‘combustionanalysis’ is used in its specific sense.

In many test cells in the world, throughout the 1990s and up to the presentdate, combustion analysis equipment may be a temporary visitor; brought in whenrequired and operated in parallel with the test bed automation system. Temporaryinstallations may give problems and inconsistencies due to signal interference andit is now common to find dedicated engine calibration cells with engine indicatingequipment as a permanent part of an optimized installation.

Currently there is no ASAM standard for the storage of raw CA data, so they tendto be stored within structured binary files designed by the makers of the analysisequipment.

There may be fewer than 10 specialist suppliers of complete IE hardware andsoftware suites worldwide and probably only half these are also specialist CA trans-ducer manufacturers. Any search of the internet will result in discovery of manyuniversities and private test-houses that have assembled their own analysis systemsfrom commercially available components.

If the instrumentation and software suites of two or more subsystems are fromdifferent suppliers, it is vital that the communications protocols are mutually under-stood and that one competent party is responsible for integration. Any engine controlunit communication link would normally be designed to an industrial standard, suchas the current ASAP-3, but third party testers may find that such outputs as alarmcodes require manufacturers’ documentation that is not in the public domain.

The combustion process and combustion analysis 283

To fully exploit the capabilities of the leading engine calibration systems, thecells are often set up to run unmanned for prolonged periods. The system alarms andinterlocks, including remote alarm indication, must be carefully considered for suchcells.

As with many other aspects of engine testing, the correlation of results producedon different systems, some of which are not well integrated, ranges from difficult toself-deluding.

Fundamental influences on combustion

At a fundamental level, the performance of an internal combustion engine is largelydetermined by the events taking place in the combustion chamber and engine cylinder.These are influenced by a large number of factors:

• configuration of the combustion chamber and cylinder head;• flow pattern (swirl and turbulence) of charge entering the cylinder, in turn deter-

mined by design of induction tract and inlet valve size, shape, location, lift andtiming;

• ignition and injection timing, spark plug position and characteristics, injector andinjection pump design, location of injector;

• compression ratio;• air/fuel ratio;• fuel properties;• mixture preparation;• exhaust gas recirculation (EGR);• degree of cooling of chamber walls, piston and bore.

For those readers not familiar with the details of the combustion process of the sparkignition (SI) or compression ignition (CI) engine, they are summarized below.

Combustion in the gasoline engine

1. A mixture of gas and vapour is formed, either in the induction tract, in whichthe fuel is introduced either by a carburettor or injector, or in the cylinder in thecase of port injection.

2. Combustion is initiated at one or more locations by an electric spark.3. After a delay, the control of which still presents problems, a flame is propagated

through the combustible mixture at a rate determined, among other factors, bythe air motion in the cylinder.

4. Heat is released progressively with a consequent increase in temperature andpressure. During this process the bulk properties of the fluid change as it istransformed from a mixture of air and fuel to a volume of combustion products.

284 Engine Testing

Undesirable effects, such as preignition, excessive rates of pressure rise, lateburning and detonation or ‘knock’ may be present.

5. Heat is transferred by radiation and convection to the surroundings.6. Mechanical work is performed by expansion of the products of combustion.

The development of the combustion chamber, inlet and exhaust passages and fuelsupply system of a new engine involve a vast amount of experimental work, someon flow rigs that model the geometry of the engine, most of it on the test bed.

Combustion in the diesel engine

1. Air is drawn into the cylinder, without throttling, but frequently with pressurecharging. Compression ratios range from about 14:1 to 22:1, depending on thedegree of supercharge, resulting in compression pressures in the range 40–60bar and temperatures from 700 to 900�C.

2. Fuel is injected at pressures that have increased in recent years from 600 to 1500bar or higher. Charge air temperature is well above the autoignition point of thefuel. Fuel droplets vaporize, forming a combustible mixture which ignites aftera delay which is a function of charge air pressure and temperature, droplet sizeand fuel ignition quality (cetane number).

3. Fuel subsequently injected is ignited immediately and the progress of combustionand pressure rise is to some extent controlled by the rate of injection. Air motionin the combustion chamber is organized to bring unburned air continually intothe path of the fuel jet.

4. Heat is transferred by radiation and convection to the surroundings.5. Mechanical work is performed by expansion of the products of combustion.

Choice of combustion ‘profile’ involves a number of compromises. A high maximumpressure has a favourable effect on fuel consumption, but increases NOx emissions,while a reduction in maximum pressure, brought about by retarding the combustionprocess, results in increased particulate emissions. Maximum combustion pressurescan exceed 200 bar.

Unlike the spark-ignition engine, the diesel must run with substantial excess air

to limit the production of smoke and soot. A precombustion chamber engine can runwith a lambda ratio of about 1.2 at maximum power, while a direct injection (DI)engine requires a minimum excess air factor of about 50 per cent, roughly the sameas the maximum at which a spark-ignition engine will run. It is this characteristicthat has led to the widespread use of pressure charging, which achieves a reasonablespecific output by increasing the mass of the air charge.

Large industrial or marine engines invariably use direct injection; in the case ofvehicle engines, the indirect injection or prechamber engine is being abandoned infavour of direct injection because of the better fuel consumption and cold startingperformance of the DI engine. Compact very high pressure injectors and electroniccontrol of the injection process have made this development possible.


Effects on combustion process of air/fuel ratio

Anyone who attempts to ‘tune’ a gasoline engine must have a clear understandingof this subject to make any progress with the task.

A gasoline engine is capable of operating on a range of air-to-fuel ratios by weightfrom about 81 to 201, and weaker than this in the case of stratified charge andlean-burn engines. Several definitions are important:

• Mixture strength. A loose term usually identified with air/fuel ratio and describedas ‘weak’ (excess air) or ‘rich’ (excess fuel).

• Air/fuel ratio. Mass of air in charge to mass of fuel.• Stoichiometric air/fuel ratio (sometimes known as ‘correct’ air/fuel ratio). The

ratio at which there is exactly enough oxygen present for complete combustionof the fuel. Most gasolines lie within the range 14:1 to 15:1 and a ratio of 14.51:1may be used as a rule of thumb. Alcohols, which contain oxygen in their make-up,have a much lower value in the range 7:1 to 9:1.

• Excess air factor or ‘lambda’(�) ratio. The ratio of actual to stoichiometricair/fuel ratio. The range is from about 0.6 (rich) to 1.5 (weak). Lambda ratio hasa great influence on power, fuel consumption and emissions.

• The equivalence ratio, �, is the reciprocal of the lambda ratio and is preferredby some authors.

A basic test is to vary the air/fuel ratio over the whole range at which the engineis able to run, keeping throttle opening and speed constant. The results are oftenpresented in the form of a ‘hook curve’, Fig. 14.1, which shows the relation betweenspecific fuel consumption and b.m.e.p. over the full range of mixture strengths.

If such a test is carried out on an engine fitted with a quartz observation windowthe following changes are observed:

• At mixture strength corresponding to maximum power and over a range of weakermixtures combustion takes place smoothly and rapidly with a blue flame whichis extinguished fairly early in the expansion stroke.

• With further weakening combustion becomes uneven and persists throughout theexpansion stroke. ‘Popping back’ into the induction manifold may occur.

• As we proceed towards richer mixtures, the combustion takes on a yellow colour,arising from incandescent carbon particles, and may persist until exhaust valveopening. This may lead to explosions in the exhaust system.

The following features of Fig. 14.1 call for comment:

• Point a corresponds to the weakest mixture at which the engine will run. Poweris much reduced and specific fuel consumption can be as much as twice thatcorresponding to best efficiency.

286 Engine Testing

Specific

fuel consum

ption (

litre

/kW

h)

1.4

1.2

1

0.8

0.4

0.2

0.6

0

b.m.e.p. (kN/m2)

Weak

Rich

Full throttle

e

a

b

d

c

0 100 200 300 400 500 600

Figure 14.1 Hook curve for gasoline engine

• Point b corresponds to the best performance of the engine (maximum thermalefficiency). The power output is about 95 per cent of that corresponding tomaximum power.

• Point c corresponds to the stoichiometric ratio.• Point d gives maximum power, but the specific consumption is about 10 per

cent greater than at the point of best efficiency. It will be evident that a primerequirement for the engine management system must be to operate at point b,except when maximum power is demanded.

• Point e is the maximum mixture strength at which the engine will run.

It is possible to produce similar curves for the whole range of throttle positions andspeeds and hence to derive a complete map of optimum air and fuel flow rates asone of the bases for development of the engine management system, whether it be atraditional carburettor or a computer-controlled injection system.

If at the same time that the hook curve is produced the air flow rate is measured,the same information may be presented in the form of curves of power output andspecific consumption against air/fuel ratio or �, see Fig. 14.2, which corresponds toFig. 14.1.


Specific

consum

ption (

litre

/kW

h)

1.5

1.0

0.5

0

Air/fuel ratio

8 9 10 11 12 13 14 16 17 18 1915

b.m

.e.p

. (k

N/m

2)

0

200

400

600

b.m.e.p.

Specific consumption

Rich Weak

Maximumpower

Minimumconsumption

Figure 14.2 Variation of power output and specific fuel consumption withair/fuel ratio

Cylinder pressure diagrams

Many special techniques have been developed for the study of the combustionprocess. The oldest is direct observation of the process of flame propagation, usinghigh-speed photography through a quartz window. More recent developments includethe use of flame ionization detectors (FID) to monitor the passage of the flame andthe proportion of the fuel burned, also hot-wire and laser Doppler anemometry.

The standard tool for the study of the combustion process is the cylinder pressureindicator; the different attributes of which are described later in this chapter. In addi-tion to cylinder pressure, a variety of quantities must be measured, using appropriatetransducers, either on a time base or synchronized with crankshaft position. Theymay include

• fuel line pressure and needle lift;• cylinder pressure;• ionization signals;• crank angle;• time;• ignition system events;

288 Engine Testing

• inlet and exhaust pressures;• various (quasi-static) temperatures.

Each signal calls for individual treatment and appropriate recording methods. Dataacquisition rates have increased in recent years and systems are available with asampling rate of up to 1MHz on 16 channels. This is equivalent to cylinder pressurestaken at 0.1� intervals at 16 000 rev/mm.

Combustion analysis aims at an understanding of all features of the process, inparticular of the profile of heat release. The test engineer concerned with enginedevelopment should be familiar with both the principles involved and the considerableproblems of interpretation of results that arise. A comprehensive account of thetheory of combustion analysis would much exceed the scope of this book, but adescription of the essential features follows.

The aim of combustion analysis is to produce a curve relating mass fraction burned(in the case of a spark ignition engine) or cumulative heat release (in the case of adiesel engine) to time or crank angle. Derived quantities of interest include rate ofburning or heat release per degree crank angle and analysis of heat flows based onthe first law of thermodynamics.

Stone and Green-Armytage1 describe a simplified technique for deriving the burnrate curve from the indicator diagram using the classical method of Rassweilerand Withrow.2 This takes as the starting point a consideration of the process ofcombustion in a constant volume bomb calorimeter. Figure 14.3 shows a curve ofpressure and rate of change of pressure against time. It is then assumed that massfraction burned and cumulative heat release, at any stage of the process, are directlyproportional to the pressure rise.

Combustion in the engine is assumed to follow a similar course, the differencebeing that the process does not take place at constant volume. There are three differenteffects to be taken into account:

• pressure changes due to combustion;• pressure changes due to changes in volume;• pressure changes arising from heat transfer to or from the containing surfaces.

Figure 14.4 shows a pressure–crank angle diagram and indicates the procedure. Thestarting point is the ‘motored’ or ‘no burn’ curve, shown dotted. This curve used to beobtained by interrupting the ignition or injection for a single cycle and recording thecorresponding pressure diagram; certain engine indicators are able to do this. Withthe more common availability of motoring dynamometers it is possible to obtainthe curve by motoring the unfired engine. An alternative, less accurate, method isto fit a polytropic compression line to the compression curve prior to the start ofcombustion and to extrapolate this on the assumption that the polytropic index nc inthe expression pvnc = constant remains unchanged throughout the remainder of thecompression stroke.


1

0.8

0.6

0.4

0.2

0

Time (arbitrary units)

Burn rate

Mass fractionburned

0 1 2 3 4 65 7 8

Figure 14.3 Combustion in a constant volume bomb calorimeter

Pressure (absolute)n ∼ 0.7

n ∼ 1

n = ne

n steadily rises

pv n

e = constant

pv n

c = constant

v1

v2

v = v0

p1

p2

A B DC t.d.c. Crank angle

Figure 14.4 Pressure–crank angle diagram showing derivation of mass fractionburned

290 Engine Testing

If A and B are two points on the compression line at which volumes and (absolute)pressures are respectively v1, v2 and p1, p2, then the index of compression is given by:

nc =log�p1/p2�

log�v2/v1�(1)

The value of the index nc during compression is likely to be in the region of 1.3, to becompared with the value of 1.4 for the adiabatic compression of air. The lower valueis the result of heat losses to the cylinder walls and, in the case of spark ignitionengines, to the heat absorbed by the vaporization of the fuel. The start of combustionis reasonably well defined by point C, at which the two curves start to diverge.

It is now necessary to determine the point D, Fig. 14.4, at which combustion maybe deemed to be complete. Various methods may be used, but perhaps the mostpractical one is a variation on the method described above for determining the startof combustion. Choose two points on the expansion line sufficiently late in the strokefor it to be reasonable to assume that combustion is complete but before exhaustvalve opening: 90� and 135� after t.d.c. may be a reasonable choice. The polytropicindex of expansion (in the absence of combustion) ne is calculated from eq. (1). Thisis also likely to be in the region of 1.3, except in the special case when burningcontinues right up to exhaust valve opening, as in a spark-ignition engine burning aweak mixture.

The next step is to calculate the polytropic index of compression/expansion forsuccessive intervals, typically one degree of crank angle, from the start of combustion.Once again eq. (1) is used, inserting the pressures and volumes at the beginningand end of each interval. The value of n varies widely as combustion proceeds, buttowards the end of the process it converges on the value ne determined above.

The point D at which the two indices become equal is generally ill-defined, asthere is likely to be considerable scatter in the values derived for successive intervals,but fortunately the shape of the burn rate curve is not very sensitive to the positionchosen for point D.

The curve of heat release or cumulative mass fraction burned is derived as follows.Figure 14.5a shows diagrammatically an element of the indicator diagram beforet.d.c. but after the start of combustion. During this interval the pressure rises fromp1 to p2, while the pressure rise corresponding to the (no combustion) index nc isfrom p1 to p0. It is then assumed that the difference between p2 and p0 represents thepressure rise due to combustion. It is given by:

�p= p2−p0 = p2−p1

(

v1

v2

)nc

(2)

It is now necessary to refer this pressure rise to some constant volume, usually takenas v0, the volume of the combustion chamber at t.d.c.. It is assumed that the pressurerise is inversely proportional to the volume:

�pc = �p�v1+v2�

2v0(3)


p1p1

p0

p0

p2

p2

v2 v1

v2

(b)(a)

v1

Figure 14.5 Derivation of mass fraction burned: (a) compression; (b)expansion

Essentially the same procedure is followed during the expansion process, Fig. 14.5b,except that here there is a pressure fall corresponding to the (no combustion) indexof expansion ne. �p and �pe are calculated from eqs (2) and (3) as before.

Finally �pc is summed for the whole combustion period, and the resultingcurve, Fig. 14.6, is taken to represent the relation between mass fraction burnedor cumulative heat release and crank angle. The ‘tail’ of the curve, following theend of combustion, point D, will be horizontal if the correct value of ne has beenassumed. If ne has been chosen too low, it will slope downwards and if too highupwards.

As with much combustion analysis work, there are a number of assumptionsimplicit in this method, but these do not seriously affect its value as a developmenttool. The technique is particularly valuable in the case of the DI diesel engine,in which it is found that the formation of exhaust emissions is very sensitive tothe course of the combustion process, which may be controlled to some extent bychanging injection characteristics and air motion in the cylinder. Figure 14.7 fromRef. 3 shows the time scale of a typical diesel combustion process.

Figure 14.8 shows an indicator diagram taken from a small single cylinder dieselengine and the corresponding heat release curve derived by means of a computerprogram modelled on the method of calculation described above.

The start of combustion, at 9� before t.d.c., is clearly defined and lags the start ofinjection by 6�. Combustion is complete by about 60� after t.d.c. and the derived valueof the (no combustion) index of expansion in this case was 1.36 (smaller enginesgenerally display a higher index, owing to the higher heat losses in proportion to

292 Engine Testing

1

0.8

Massfractionburned

Cumulativeheatrelease

0.6

0.4

0.2

0

t.d.c. Crank angle

Figure 14.6 Mass fraction burned plotted against crank angle

volume). Figure 14.9 shows a similar trace taken from a larger automotive diesel bya combustion analysis instrument.

Total and instantaneous energy release

Many of the required results used by the engine developer are derived from thefundamental calculations relating to energy release in the cylinder. This can beexpressed as total energy per cycle (expressed as a pressure value × m.e.p.) or inthe calculation of instantaneous energy release with respect to crank angle (either asa rate or integral curve).

These calculations are well known and supported by established theory. Experi-menters may adjust or adapt the calculations according to engine type or application.Any assumptions made in these calculations, in order to simplify them and reducecalculation time, can be adjusted according to preference and laboratory procedure.However, this can lead to the analysis becoming somewhat subjective, it is thereforeimportant to know which theory and calculation method has been adopted whencomparing results.

In this section, the basic principle behind these calculations will be discussed forthe purpose of an overview; established publications shown in the reference sectioncan be used as a source if more detailed information is needed.


Start of fuel pump actionStart of injection

End of injection

End of effective combustion

End of effective air entrainment

Crank angle

Time(ms at 250 rev/min)

(ms at 2500 rev/min)

(ms at 5000 rev/min)

Atomi zation

Start of combustion

Peak pressure

a injectionb evaporationc air entrainment and mixingd heat releasee heat transferf NOx formation

g soot and CO formation

Peak temperature

f

0 15 30 45 60 75

5040302010

1

0.5 1.0 1.5 2.0 2.5

2 3 4 5

d

cb

a

eg

t.d.c.

Figure 14.7 Time scale of diesel combustion process

Cyclic energy release – mean effective pressure (indicated,gross and pumping)

Cylinder pressure data can be used to calculate the work transfer from the gas to thepiston. This is generally expressed as the indicated mean effective pressure (i.m.e.p.)and is a measure of the work output for the swept volume of the engine. The resultis a fundamental parameter for determining engine efficiency as it is independent ofspeed, number of cylinders and displacement of the engine.

The i.m.e.p. calculation is basically the enclosed area of the high pressure part ofthe pV diagram and can be derived via integration:

i�m�e�p� =∫

P�dV

This effectively gives energy released or gross m.e.p. (gross work over the compres-sion and expansion cycle, g.m.e.p.). Integration of the low pressure (or gas exchange)

294 Engine Testing

80

60

40

20

0

–20

–180 –160 –140 –120 –100 –80 –60 –40 –20


Cylinder pressure

Heat release

Heat re

lease (%

)C

ylin

der

pre

ssure

(bar)

0 20 40 60 80 100 120 140 160 180

10

9

8

7

6

5 100

80

60

40

20

0

4

3

2

1

0

–1

–2

–3

–4

–5

–6

–7

–8

–9

–10

Figure 14.8 Cylinder pressure diagram and heat release curve for a small singlecylinder diesel engine

Inte

gra

ted (

%)

Insta

nta

neous (

%/d

egre

e)

125

100

75

50

25

0

6

8

4

2

0

–2

–20 0 20 40 60 80


Figure 14.9 Diagram showing instantaneous and integrated energy release vscrank angle. (Source: AVL)


part of the cycle gives the work lost during this part of the process (pumping lossesalso known as p.m.e.p.). Subtraction of this value from the gross i.m.e.p. gives thenet i.m.e.p. (n.m.e.p.) or actual work per cycle.

Therefore, it can be stated that:

n�m�e�p� = g�m�e�p� − p�m�e�p�

The most important factor to consider when measuring i.m.e.p. is that the t.d.c.position (see section ‘Exact determination of true top dead centre position’ below).

In order to optimize the calculation note that it is not necessary to acquire data athigh resolution. Measurement resolution of a maximum of one degree crank angleis sufficient, higher than this does not improve accuracy and is a waste of systemresources.

Also important for accurate i.m.e.p. calculation are the transducer properties;mounting location and stability of the transducer sensitivity during the engine cycle.

Instantaneous energy release

The cylinder pressure data, in conjunction with cylinder volume can be used toextrapolate the instantaneous energy release from the cylinder with respect to crankangle. There are many well-documented theories and proposals for this, but all arefundamentally similar and stem from the original work carried out by Rassweilerand Withrow.2

The calculations can range from the quite simple with many assumptions, to verysophisticated simulation models with many variables to be defined and boundaryconditions to be set.

The fundamental principle to the calculation of energy release relies on the defini-tion of the compression and expansion process via a polytropic exponent (ratio of spe-cific heats under constant pressure and volume conditions for a given working fluid).

If this can be stated with accuracy then a motored pressure curve can be extrap-olated for the fired cycle. The fired curve can then be compared with the motoredcurve and the difference in pressure between them, at each angular position, and withrespect to cylinder volume at that position allows calculation of energy release.

A typical simplified algorithm to calculate energy release:

Qi =K

�−1

[

� ·pi · �Vi+n−Vi−n�+Vi · �pi+n−pi−n�]

(4)

This calculation can be executed quickly within a computerized data acquisitionsystem, but it does not account for wall or blow-by losses in the cylinder. Also, itassumes that the polytropic index is constant throughout the process. These compro-mises can affect accuracy in absolute terms but in relative measurement applications

296 Engine Testing

this simplified calculation is well proven and established in daily use. Under con-ditions where engine development times are short, algorithms have to be simplifiedsuch that they can be executed quickly.

Most importantly for the engine developer are the results which can be extractedfrom the curves which give important information about the progress and quality ofcombustion.

Computerized engine indicating technology and methodology

The modern engine indicating or combustion analyser creates data that are largerby at least an order of magnitude when compared with the log point data measuredand stored by a typical test cell automation system. Therefore, appropriate method-ology for handling, reducing and analysing these data is an essential function of EIequipment.

The combustion analysis device is the reciprocal of the test bed with respect toraw and calculated data. The ratio of formula to channels is far higher for a com-bustion analyser and an efficient way of creating and standardizing thermodynamiccalculations is another important part of the overall efficiency of the process.

Key to the whole process is the measurement of pressure within the combustionchamber during the complete four-stroke or two-stoke process. This process in exactdetail can be highly variable between cylinders and even within cycles inside the samecylinder hence the need for high levels of data analysis to enable mean performancelevels to be determined.

The tools used to measure cylinder pressure are pressure transducers that work onthe piezoelectric principle, wherein the change of pressure creates a change in circuitcharge. The circuit itself is quite unlike any other sensor or transducer commonly seenin an engine test environment and is therefore important that users fully understandthe special handling requirements of this technique.

Basic circuit and operation of pressure measurement chain

Piezoelectric combustion pressure sensors are always used with appropriate sig-nal conditioning in combustion measurement applications. The signal conditioningrequired for a charge signal is a charge amplifier and this basically consists of ahigh-gain inverting voltage amplifier with a MOSFET (metal–oxide–semiconductorfield-effect transistor) or JFET (junction field-effect transistor) at its input to achievehigh insulation resistance.

The purpose of the charge amplifier is to convert the high impedance chargeinput into a usable output voltage. The basic circuit diagram is shown in Fig. 14.10.A charge amplifier consists of a high gain amplifier and a negative feedback capacitor(CG). When a charge is delivered from a piezoelectric pressure transducer (PT), thereis a slight voltage increase at the input of the amplifier (A). This increase appears


Long

ShortReset

Charge amplifier

VO

RINS

IINS

CF

CC

PT

UI

A

+

+

–

–

II

I

RF

Figure 14.10 Basic charge amplifier circuit (AVL List)

at the output substantially amplified and inverted. The negatively biased negativefeedback capacitor (CF) correspondingly taps charge from the input and keeps thevoltage rise small at the amplifier input.

At the output of the amplifier (A), the voltage (VO) sets itself so that it picks upenough charge through the capacitor to allow the remaining input voltage to resultin exactly VO when amplified by A. Because the gain factor of A is very large (upto about 100 000), the input voltage remains virtually zero. The charge output fromthe pressure transducer is not used to increase the voltage at the input capacitances,it is drawn off by the feedback capacitor.

With sufficiently high open-loop gain, the cable and sensor capacitance can beneglected; therefore, changes in the input capacitance due to different cables withdifferent cable capacitance (CC) have virtually no effect on the measurement result.This leaves the output voltage dependent only on the input charge and the rangecapacitance. That is,

VO = Q/CF

An output is produced only when a change in state is experienced, thus the piezo-electric transducer and charge amplifier cannot perform true static measurements.This is not an issue in normal combustion measurements, as they require a highlydynamic measurement chain to adequately capture all aspects of the phenomenon.Special considerations are required for calibration of the measurement chain and indata analysis.

While the modern charge amplifier measurement chain, Fig. 14.11, is a robustand well-proven measurement and the technique is almost universally adopted for

298 Engine Testing

Storage of data onlocal or global location

PC and user interface

Amplifier

Combustion pressuresensor

Indicatingsystem

ADC

Digitalsignal processor

PC-system

Data post-processing

Amplifiers

Signal/sensor

Figure 14.11 Typical workflow process through combustion analysis

combustion pressure measurement, there are important issues and measurement char-acteristics to consider:

• Drift. This is defined as an undesirable change in output signal over time thatis not a function of the measured variable. In the piezoelectric measurementsystem, there is always an inherent drift in the output signal due to the workingprinciple of the system. Electrical drift in a charge amplifier can be caused bylow insulation resistance at the input.

• Time constant. The time constant of a charge amplifier is determined by theproduct of the range capacitor and the time constant resistor. It is an importantfactor for assessing the capability of a piezoelectric measurement system formeasurement of very slow engine phenomena (cranking speed) without any sig-nificant errors due to the discharge of the capacitor. Many charge amplifiers haveselectable time constants that are altered by changing the time constant resistor.

Drift and time constant simultaneously affect a charge amplifier’s output. One or theother will be dominant. There are a number of methods available in modern amplifiertechnology to counteract electrical drift and modern charge amplifier technologycontains an electronic drift compensation circuit.

If it is possible to maintain extremely high insulation values at the pressuretransducer, amplifier input, cabling and associated connections, leakage currents canbe minimized and high quality measurements can be made. In order to do thisthough, the equipment must be kept clean and free of dirt and grease to laboratorystandards. In practice, a normal engine test cell environment does not provide thislevel of cleanliness and this method of preventing drift is not a practical proposition.Operation of the amplifier in short time constant mode means that the drift due to


input offset voltage can be limited to a certain value and thus drifting into saturationcan be prevented.

• Filtering. Electrical filtering is generally provided in the charge amplifier toeliminate certain frequencies from the raw measured data. Typically high passfilters are used to remove unwanted lower frequency components; for combustionmeasurement, this may be necessary in a knock measurement situation whereonly the high frequency components are of interest.

The high pass filter will allow optimization of the input range on the measurementsystem to ensure the best possible analogue to digital conversion of the signal. Lowpass filters can be implemented to remove unwanted high frequency, interferencesignal content from the measurement signal such as structure-borne noise signalsfrom the engines that are transmitted to the transducer.

It is an important point to bear in mind when using electrical filters that a certainphase shift will always occur and that this can cause errors which must be considered.For example, any phase shift has a negative effect on the accuracy of the i.m.e.p.determination when a low pass filter is used. The higher the engine speed, the higherthe lowest permitted filter frequency. As a general rule, in order to avoid unacceptablephase shift, the main frequency of the cylinder pressure signal should not be morethan 1 per cent of the filter frequency.

Ground loops in indication equipment

These are discussed in Chapter 10 and can be a significant problem for piezoelectricmeasurement chains (due to the low level signal) and can be extremely difficult todetect. It is therefore prudent to take appropriate precautions to prevent them andhence avoid their effects. The simplest way is to ensure all connected componentsare at a common ground via interconnection with an appropriate cable; this can bedifficult to implement if equipment is not permanently installed at the test cell.

Ground isolated transducers and charge amplifiers can be purchased that willeliminate the problem, but it is important to note that only one or the other of thesesolutions should be used. If used together the charge circuit will not be complete andhence will not function!

Modern amplifier technology utilizes developments in digital electronics and soft-ware to correct or allow correction of signal changes occurring during the test period.The amplifier settings are software driven via the instrumentation PC, which meansthat settings can be adjusted during measurements without entering the test cell.Local intelligence at the amplifier provides automatic gain factor calculation fromsensor sensitivity, pressure range and output voltage thus providing optimum adap-tation of the input signal with nearly infinitely variable gain factors. This optimizinginformation can be stored on board the amplifier complete with sensor parameters.

Modern amplifiers are physically small which allows mounting very close tothe actual sensors and this means short cables between sensors and amplifier and

300 Engine Testing

therefore the best protection from electromagnetic interference. The direct connectionof piezo sensors without interconnection reduces signal drift and possible problemswith grease and humidity on intermediate connectors.

Exact determination of true top dead centre position

The test engineer in possession of modern equipment continues to be faced withwhat appears to be a rather prosaic problem: the determination of the engine’s topdead centre point.

This is a more serious difficulty than may be at first apparent. An electronicengine indicator records cylinder pressure in terms of crank angle, and measurementsat each 1/10 of one degree of rotation are commonly available. However, in orderto compute indicated power it is necessary to transform the cylinder pressure-crankangle data to a basis of cylinder pressure-piston stroke. This demands a very accuratedetermination of crank angle at the top dead centre position of the piston.

If the indicator records top dead centre 1� of crankshaft rotation ahead of the trueposition the computed i.m.e.p. will be up to 5 per cent greater than its true value.If the indicator records t.d.c. 1� late computed i.m.e.p. will be up to 5 per cent lessthan the true value.

Top dead centre can now be accurately determined, for a particular running con-dition, by using a special capacitance sensor such as the AVL 428 unit, which ismounted in a spark plug hole of one cylinder while running the engine on the remain-ing cylinders. The t.d.c. position thus determined is then electronically recorded byusing the angle of rotation between t.d.c. and a reference pulse using a rotationalencoder attached to the crankshaft. Another technique makes use of recording thecylinder pressure peak from a cylinder pressure transducer but without such specialtransducers and the required software to match the shaft angle, precise determinationof geometrical t.d.c. is not easy. The usual and time-honoured method is to rotatethe engine to positions equally spaced on either side of t.d.c., using a dial gauge toset the piston height, and to bisect the distance between these points. The rotationposition of this even then had to be fixed on the flywheel in such a way as to berecognized by a pick-up.

This traditional method gives rise to several sources of error:

• the difficulty of carrying out this operation with sufficient accuracy;• the difficulty of ensuring that the signal from the pick-up, when running at speed,

coincides with the geometrical coincidence of pin and pick-up;• torsional deflections of the crankshaft, which are always appreciable and are

likely to result in discrepancies in the position of t.d.c. when the engine is running,particularly at cylinders remote from the flywheel.

In high performance engines, the t.d.c. position does not remain entirely constantduring the test run due to physical changes in shape of the components; in thesecases a t.d.c. position specific to a particular steady state may have to be calculated.


Integration of engine indication (EI) equipment within thetest cell

Whatever the level of sophistication the physical equipment involved in enginecalibration and EI work can be considered as a high speed data acquisition, storage,manipulation and display system. It will consist of three subsystems:

• Engine rigged transducers: some of which may require special engine preparation,such as machining of heads for pressure transducers and mounting of an opticalspeed encoder.

• Interconnection loom which, in the case of mobile systems, will have to passthrough the firewall between cell and control room.

• Data acquisition and analysis equipment mounted in, and fully integrated with,the control room instrumentation.

High end engine calibration cells

Let us consider the operational problem faced by the cell operators of an OEM testcell carrying out the optimization of an engine control map. There are three distinctsubsystems producing and displaying data while the engine is running.

The test cell automation system that is controlling the engine performance andmonitoring and protecting, via alarm logic, both engine and test facility operation.The engine ECU that is controlling the fuelling and ignition using engine-mountedparameters to calculate essential variables that it is incapable of measuring directlysuch as torque from inlet manifold pressure.

The combustion analysis system that is using special (non-vehicular) transducersto directly measure cylinder pressure against crank shaft rotation and is able to deriveand display such values as

• mean effective pressures, i.m.e.p., p.m.e.p., etc.;• location of peak pressure and pressure rise rates;• mass faction burn parameters;• polytropic index;• injector needle movement and injected fuel mass;• heat release, etc.

To enable such instrumentation and data to efficiently fulfil its purpose in the min-imum time, the operator of the cell needs to be highly skilled and experienced andvery careful consideration needs to be given to the design of the experiments (DoE)being carried out on the test bed.

The task described is now so common that often there are insufficient highlyskilled and experienced operators available. In response, the test equipment industryhas produced integrated suites of instrumentation and software that, to a degree, canreduce the skill levels required to design the test sequences and to efficiently locatethe boundary operating conditions of the test engine (see Chapter 19).

302 Engine Testing

bar

50

40

30

20

10

0

deg CA

–60 –45 –30 –15 0 15 30 45 600

0.5

1.0

mm200

400

600

800

1000

bar

Injectionlinepressure

Nozzleneedlelift

Lin

e p

ressure

Cylin

der

pre

ssure

Needle

lift

Combustionchamberpressure

Figure 14.12 Typical engine indicating display parameters

Cells that carry out engine calibration work such as that described above have tobe designed from the beginning as an integrated system if they are to work efficiently.All of the data having different and various native abscissa, such as time base, crankangle base and engine cycle base, communicated via different software interfaces(COM, ASAP, CAN bus, etc.), have to be correctly aligned, processed, displayed andstored which is no mean feat. In fully automated beds, it is common to have upwardsof six 17-inch monitors on the operator’s desk. In addition to the pressure transducersrequired for cylinder pressure, additional transducers are required to measure eventsand consequences of the combustion that occurs on the same timescale, such asinjector operation and fuel rail pressure (see Fig. 14.12).

Middle market, occasional combustion analysis cells

The operator of and engine test required to carry out occasional map modificationwork performance testing will find that much of the combustion analysis equipmentis often, although not always, roving trolley-mounted plant rather than a permanentfixture in the cell and control room. This means that unlike exhaust gas emission plant,the only real space requirement imposed is that of the EI equipment operator; thisis because such equipment is invariably operated by, and in the care of, specialists.Test sequences designed to support IE work are, at this level of work, usually createdwith off-line software or problem areas of engine performance are located throughmanual control.

Control cables fitted to the ECU via large multiway plugs give practical riggingproblems when having to be passed through the cell/control fire wall.


Typical engine indicating plant will have a ‘pseudo-oscilloscope’ display whichwill be separate from the default cell control system display, therefore desk spacefor two operators and an absolute minimum of two flat display screens should becatered for.

‘Knock’ sensing

One of the most important boundary conditions that an engine control system mustbe programmed to avoid is ‘knock’. This destructive, spontaneous ignition of unburntgas in the cylinder can be detected in the engine test cell by

• characteristic pressure pulse signature; the normal method used by EI equipmentusing specially rigged engines;

• an accelerometer tuned to look for the vibration caused by knock; the commonmethod used by ECU since the transducer is appropriate for use in productionvehicles;

• microphone and operator detection is still a valid method but probably onlyappropriate to specialist motor sport test facilities where low engine numbers andrapidly changing ECU maps do not allow for standard models to be formed.

Engine indicating pressure transducers (EIPT)

Cylinder pressure measurements play the key role in any engine indicating andcombustion analysis work. The choice of pressure transducer from the wide rangenow commercially available, its correct mounting in the engine and its integrationwithin the calibrated system are vital ingredients in obtaining optimum results.

Up to the early 1990s commercially available piezoelectric pressure transducerswere more temperature sensitive both in terms of cyclic effects and absolute max-imum working temperature than units available today. The temperature effects aredue to exposure to the combustion process and most would fail if required to oper-ate above 200�C, measured at the diaphragm. Consequently, most engine indicatingpressure transducers (EIPTs) were water cooled. Water-cooled EIPTs will operate inmost i.c. engine combustion chambers, including turbocharged units, but it is vitalthat the cooling system meets the following conditions:

• The transducer cooling system, while in use, must be integrated with the controlsystem shutdown circuit so as to ensure that the cooling is running before enginestart and during the complete termination of testing and cooldown of the engine.Transient dips in supply pressure will risk permanent loss of the transducers.

• The water should be distilled (deionized) and filtered. The transducers have verysmall passages that will become ineffective or blocked by scale.

304 Engine Testing

• The deionized water should be supplied at as low and constant pressure as possibleto ensure flow, so as to avoid changes in internal transducer pressure that couldcorrupt output signal.

Uncooled EIPTs are being developed continually and currently will operate to over400�C which covers the majority of naturally aspirated automotive work outsidespecialist and motor-sport development work.

Mounting of the EIPT and special designs

EIPTs are made in various sizes being threaded at the diaphragm end in sizes rangingfrom 6 up to 18mm. They are normally inserted into the combustion chamber sothat their tips are flush with the parent combustion chamber material. This gives riseto problems of finding space and material in the engine head structure to support thetransducer, so special transducers which take the form of an instrumented spark plugfor gasoline engines, or glow plugs for diesel engines, have been developed.

Speed/crank angle sensors

For all tests involving engine indicating work, the engine will have to be rigged atits free crank-shaft end with an incremental optical shaft encoder. The alignmentof the drive of these devices has to be undertaken with care and using the correctcomponents to suit the application. The devices consist of a static pick-up that isable to read positional data from an engine shaft-mounted disk that is etched withtypically 3600 equally spaced lines producing that number of square wave pulsesper revolution in the device output. There may be a number of missing lines whichenable a discrete angular position to be established as an event trigger. The disksare normally made of a glass material, except in cases where very high shock forcesare present where steel disks with lower pulse counts may be required. Some crankangle encoders have a second output that determines direction of rotation.

Result calculations for combustion analysis

Calculation of critical results from the large store of raw combustion data is an essen-tial part of the analysis and forms the basis of intelligent data reduction. Result calcu-lation and a detailed analysis of the combustion process can be performed off-line, butthis is not always convenient as the process takes time. Advances in technology allowsome systems to carry out result calculation to a high degree of accuracy on-line.

These systems can display the calculated data during measurement so that theoperator or engineer can view the data and react accordingly. Alternatively, theseresults can be transferred immediately to the test bed for data integration (into the


automation system data model) or for control purposes as in tests where the controlmode used is speed vs i.m.e.p. for cold start calibration work

Results derived from raw measured data

The actual results which are derived from the raw measured data and calculatedcurves can be grouped logically into direct and indirect results.

Direct results

These are derived directly from the raw cylinder pressure curve. They are simpleresults to calculate which can be returned quickly. Most of them can also be calculatedbefore the end of the complete engine cycle. Generally, they are less sensitive andany signal error will create a result calculation error of similar magnitude. Typicaldirect result calculations are

• maximum pressure and position;• maximum pressure rise and position;• knock detection;• misfiring;• combustion noise analysis;• injection/ignition timing;• cyclic variation of above values.

These results will be acquired from data with acquisition resolution appropriate to thetask. Resolution of pressure and pressure rise will typically be at one degree crankangle, certain values will need higher resolutions due to higher frequency componentsof interest (combustion knock or noise) or the need for accurate determination ofangular position (injection timing).

Indirect results

These results are more complex as they are derived indirectly from the raw pressurecurve data. They are reliant on additional information like cylinder volume, plus otherparameters, such as the polytropic exponent. These results are much more sensitiveto correct set-up of the equipment and the error in a signal acquired is multipliedconsiderably when passed into a calculation. Hence the result error can be greaterby an order of magnitude. Typical indirect results are

• heat release calculation (dQ, integral);• indicated mean effective pressure;• pumping losses, p.m.e.p.;• combustion temperature;

306 Engine Testing

• burn rate calculation;• friction losses;• energy conversion;• gas exchange analysis;• residual gas calculation;• cyclic variation of above values;

and commonly are interpreted as:

Result Interpretation

Mass burn fraction 0–5% Start of combustion, early flame growth

Mass burn fraction 50% Correlation with MBT for gasolineengine

Mass burn fractions 0–5% and10–90%

End of combustion, turbulent flamegrowth

Max cylinder pressure and angle Cylinder head loading, mechanical limits

g.m.e.p., n.m.e.p., p.m.e.p. Mechanical efficiency and friction losses

Coefficient of variance of i.m.e.p. Stability indicator

Most of these results, although derived through complex calculations, do not placea heavy demand on the acquisition hardware, nor do they require high resolutionswith respect to crank angle (typically one degree is sufficient) or analogue-to-digitalconversion (12–14 bit). The main demand is for sufficient calculating power toexecute the algorithms and return results and statistics as quickly as possible.

Summary

As discussed in this section, extraction of result values from the measured andcalculated curves is an important function for efficient data analysis and handling.For the combustion analysis system important features are:

• real-time or on-line calculation and display of all required results for screendisplay or transfer to the test bed via software interface or as analogue voltagevalues;

• features in the user interface for easy parameterization of result calculations, orfor definition of new, customer-specific result calculation methods without therequirement for detailed programming knowledge;

• flexibility in the software interface for integration of the combustion analyserto the test bed allowing remote control of measurement tasks, fast result datatransfer for control mode operation, tagging of data files such that test bed andcombustion data can be correlated and aligned in post processing;


• intelligent data model which allows efficient packaging and alignment of all datatypes relevant in a combustion measurement (example, crank angle data, cyclicdata, time-based data, engine parameters, etc.).

Most importantly, engine indicating and combustion analysers are always very sen-sitive to correct parameterization and the fundamentals, that is:

• correct determination of t.d.c. (particularly for i.m.e.p.);• correct definition of the polytropic exponent (particularly for heat release);• correct and appropriate method of zero level correction.

Some readers with particular interests in engine indicating and mapping may findrelevant information concerning data processing in Chapter 19.

Notation

Pressure, beginning of interval p1 bar

Pressure, end of interval p2 bar

Pressure, end of interval, no burn p0 bar

Volume, beginning of interval v1 m3

Volume, end of interval v2 m3

Volume at top dead centre v0 m3

Pressure change due to combustion �p bar

Pressure change, normalized to t.d.c. �pc bar

Polytropic index, general n

Polytropic index, compression nc

Polytropic index, expansion (no burn) ne

References

1. Stone, C.R. and Green-Armytage, D.I. (1987) Comparison of methods for the calculationof mass fraction burnt from engine pressure-time diagrams, Proc. I. Mech. E., 201 (D1).

2. Rassweiler, G.M. and Withrow, L. (1938) Motion pictures of engine flame propogation.Model for S.I. engines, S.A.E. Jl., 42, 185–204 (translation).

3. Bazari, Z. (1992) A D.I. Diesel Combustion and Emission Predictive Capability for Use in

Cycle Simulation, S.A.E. Paper No. 920462.

15 The test department organization,health and safety management,risk assessment correlation ofresults and design of experiments

Introduction

This chapter seeks to describe the administrative and technical management tasksand organization of a medium-sized test department, suggesting how the constituentparts should fit together and interact. Responsibilities imposed by health and safetylegislation and best practice are mentioned and a suggested format for risk assess-ments is included. The prime task of technical management is to ensure that the testequipment is used to its optimum efficiency which in many cases is organizationallyinterpreted as ensuring plant achieves maximum ‘up-time’. The efficiency of manytest facilities is judged by periodic cell up-time or ‘shaft rotation’ figures. But, how-ever well motivated an organization is, it can nonetheless work as ‘busy fools’ if thetasks undertaken are not organized in such a way as to be time- and cost-efficient,therefore the subject of design of experiments has been included in the chapter.

Management and group roles

Although the organizational arrangements may differ a test facility will employ staffhaving three distinctly different roles:

1. facility staff and management charged with building, maintaining, developingand the installed plant, its support services and the building fabric;

2. the internal user group charged with designing and conducting tests, collectingdata and disseminating information;

3. a quality group charged with audit to QA standards and instrument calibration.

Each group will have some responsibility for two funding streams, operational andproject specific. In both cases the first essential is to understand the brief, whichmust always include an indication of time and cost. The growing dichotomy betweenthe first two of these roles, caused by ever narrower specialization in the user group

The test department organization, health and safety management 309

and increasing scarcity of multidisciplinary facility or project staff, was the originaljustification for the first edition of this book. It is interesting to note that one ofthe tasks that may tend to ‘float’ between groups is that of specification of new ormodified facilities; referring back to Chapter 1, it is not unusual for test industrysuppliers to negotiate with a purchase or facility ‘customer’ and deliver to a ‘user’with a different requirement, the responsibility for avoiding such wasteful practicesand having a common specified requirement is that of the first level of commonmanagement.

The listed tasks are not mutually exclusive and in a small test shop may bemerged, although in all but the smallest department management of the QA task,which includes the all-important responsibility for calibration and accuracy of instru-mentation, should be kept distinct from line management of the users of the facility.

Figure 15.1 shows the allocation of tasks in a large test department and indi-cates the various areas of overlapping responsibility, also the impact of ‘qualitymanagement’. Periodical calibration of dynamometers, instruments, tools and themaintenance of calibration records may be directly in the hands of the quality man-ager or carried out by the facilities department under his supervision; either waythe procedure or computerized process must be clearly laid down. Any test facilityrequiring registration to national or international quality assurance standards willhave to provide and record a rigorous control of the calibration process for all of theinstrumentation used.

Computerized data logging has reduced the role of the test cell technician, whoused to spend much of his time reading instruments and filling in log sheets. Pressureon costs has reduced the number of technicians, who now spend more of theirtime rigging and setting up engines, while the graduate test engineer tends to spendmore of his time in the control room in the immediate consideration of the databeing generated. The profile of the staff using the control room∗ is of fundamentalimportance to the design and operation of the test department; while there is no‘correct’ solution it has to be optimized for the particular needs of each site.

Health and safety and risk management

Health and safety (H&S) matters have already been mentioned several times inconnection with test cell operation. Everyone employed or visiting a test facility hasresponsibilities (under the law in the UK) in this connection. Formal responsibilityfor H&S within a large organization will be that of a manager trained to ensure thatpolicies of the company and legal requirements are adhered to by the supervisoryorganization.

∗ The word ‘using’ should be taken in the widest sense. Users of the control area, duringnormal operation, ranges from a single production worker to a technical conference ofcustomers and sponsors. During calibration or service periods it may house internal andsubcontract specialists; the managerial control of the space needs to be clear to all of theusers.

Qualitycontrol &

audit

Facilitiesmanagement

Servicesmaintenance

Test plantmaintenance

Test runningand out-of-

hours call-out

Test-piecerigging

Test plantcalibration

Testmanagement

Data collection

Healthand

safety

Executive

management

programme control

Analysis of data

Engineeringmanagement

Report creation

Softwaresupport

Creation of testschedules and

boundaryparameters

Buildingmaintenance

Figure 15.1 Allocation of tasks within a management structure of a medium/large test facility


Risk can be defined as the danger or potential of injury, financial loss or technicalfailure. While H&S managers will concentrate on the first of these, senior managershave to consider all three at the commencement of every new testing enterpriseor task.

The legislatively approved manner of dealing with risk management is to imposea process by which, before commencement, a responsible person has to carry outand record a risk assessment. A form containing the minimum information requiredis shown in Fig. 15.2. Risk assessment is not just a single paper exercise, in alist of tasks, required by a change in work circumstances, it is a continuous task,particularly during complex projects where some risks may change by the minutebefore disappearing at task completion.

Staff involved in carrying out risk assessments need to understand that the objectof the exercise is less about describing the risk and more about putting in place

RISK ASSESSMENT

Activity: Description

Hazard Probability (P)Severity (S)Risk (Max

25)

Minimize risk by Minimizeresidual risk

by

Brief description of

specific hazard

P = 1–5S = 1–51 = rare5 = probableRisk = S × P

Description of

primary actions

required plus

any specific

responsibilities of job

holders. Reference to

attached procedures if

required.

Secondary

actions

required,

e.g. training

or

maintenance

schedules

Notes

Assessment carried out by:

Version and date:

Distribution:

Figure 15.2 A basic risk assessment form and risk ‘scoring’ method

312 Engine Testing

realistic actions and procedures in order to eliminate or reduce the potential effectsof the hazard.

Both risks of injury (acute) and risks to health (chronic), such as exposure tocarcinogenic materials, should be considered in the risk assessments.

There are important events within the life cycle of test facilities when H&Sprocesses and risk assessment should be applied:

• planning and prestart∗∗ stages of a new or modified test facility: both projectspecific and operational;

• at the change of any legislation covering explicitly or implicitly the facility;• service, repair and calibration periods by internal or subcontract staff;∗∗∗

• a significantly different test object or test routine such as those requiringunmanned running or new reference fuels;

• addition of new instrumentation.

The formal induction of new staff joining a test facility workforce and the regularreview of the levels of training required with its development are important partsof a comprehensive H&S and environmental policy. In any company that carriesout an annual appraisal of staff the subject of training will be under review by bothmanagement and staff members; where no such policy exists, training should be theformal responsibility of line management.

Planning, executing and reporting the test programme

It must never be forgotten that the sole purpose of research and development testing isto prove, or improve, the performance of the engine or component that is the subjectof the test. The sole purpose of most other tests is to check that the performance ofthe engine or complete system conforms to some specified standard.

It follows that the data assembled during the test is only of value in so far asit contributes to the achievement of these purposes. It is only too easy, now thatfacilities are so readily available for the collection of colossal quantities of data,to regard the accumulation of data as an activity of value in its own right. It isessential that the manager or engineer responsible for the operation of a test facilityshould keep this very much in mind. In the experience of the authors nothing is morehelpful in this respect than the discipline of keeping a proper log book and of writingproperly structured reports.

∗∗ In the UK under CDM regulations, the customer has responsibilities for advising thecontractors of any pre-existing conditions of the site that may affect any risk assessmentthey carry out.

∗∗∗ The provision of a risk assessment by a subcontractor does not abrogate the responsibilityof the site management for H&S matters related directly or indirectly to the work beingdone by the subcontractor. The quality of the assessment and the adherence to the processesrequired need to be checked.


The whole idea of a written log book may be considered by many to be oldfashioned in an age in which the computer has taken over. Computers, however, arenot able to report those subjective recordings of a trained technician, nor are the datathey hold always available to those who require it in the unplanned absence of thewriter.

The log book is also a vital record of all sorts of peripheral information onsuch matters as safety, maintenance, suspected faults in equipment or data recordingand, last but most important, as an immediate record of ‘hunches’ and intuitionsarising from a consideration of perhaps trivial anomalies and unexpected features ofperformance. It must be obvious that to be of real benefit the log book must be readand valued by the supervisory chain.

Test execution, analysis and reporting

The execution, analysis and reporting of a programme of tests and experiments aredifficult arts and involve a number of stages:

• First of all, the experimental engineer must understand the questions that hisexperiment is intended to answer and the requirements of the ‘customer’ who hasasked them.

• There must be an adequate understanding of the relevant theory.• The necessary apparatus and instrumentation must be assembled and, if necessary,

designed and constructed.• The experimental programme must itself be designed, with due regard to the

levels of accuracy required and with an awareness of possible pitfalls, misleadingresults and undetected sources of error.

• The test programme is executed, the engineer keeping a close watch on progress.• The test data are reduced and presented in a suitable form to the ‘customer’ and

to the level of accuracy required.• The findings are summarized and related to the questions the programme was

intended to answer.

Finally, the records of the test programme must be put together in a coherent formso that in a year’s time, when everyone concerned has forgotten the details, it willstill be possible for a reader to understand exactly what was done. Test programmesare very expensive and often throw up information the significance of which is notimmediately apparent, but which can prove to be of great value at a later date.

Formal reports may follow a similar logical sequence:

• objective of experimental programme;• essential theoretical background;• description of equipment, instrumentation and experimental method;• calculations and results;• discussion, conclusions and recommendations.

314 Engine Testing

In writing the report, the profile of the ‘customer’ must be kept in mind. Acustomer who is a client from another company will require rather different treatmentfrom one who is within the same organization. There will be common characteristics;the customer

• will be a busy person who requires a clear answer to specific questions;• will probably not require a detailed account of the equipment, but will need

a clear and accurate account of the instrumentation used and the experimentalmethods adopted;

• will be concerned with the accuracy and reliability of the results;• must be convinced by an intelligent presentation that the problem has been

understood and the correct answers given.

Although English is commonly used in the engine test industry worldwide, nativewriters of English should be aware of readers for whom English is a second or thirdlanguage and avoid idioms and, most importantly, give the full meaning of acronyms.

Cell to cell correlation

It is usual for the management of a test department to wish to be reassured that allthe test stands in the test department give the same answer, and it is not unusual foran attempt to be made to answer this question by the apparently logical procedure oftesting the same engine on all the beds. The outcome of such test based on detailedcomparison of results is invariably a disappointment, and can lead to expensive andunnecessary disputes between the test facility management and the supplier of thetest equipment.

It cannot be too strongly emphasized that an engine, however sophisticated itsmanagement system, is not suitable for duty as a standard source of torque.

In Chapter 10, we discuss the very substantial changes in engine performancethat can arise from changes in atmospheric (and hence in combustion air) conditions,in addition engine power output is highly sensitive to variations in fuel, lubricatingoil and cooling water temperature and it would be necessary to equalize these verycarefully if meaningful results were to be hoped for.

Finally, it is unlikely that a set of test cells will be totally identical: apparentlysmall differences in such factors as the layout of the ventilation air louvres and inthe exhaust system can have a significant effect on performance.

A fairly good indication of the impossibility of using an engine as a standard inthis way is contained in the standard, BS 5514, discussed below. This standard liststhe ‘permissible deviation’ in engine torque as measured repeatedly during a singletest run on a single test bed as 2 per cent. This apparently wide tolerance is no doubtbased on experience and, by implication, invalidates the use of an engine to correlatedynamometer performance.

There is no substitute for the careful and regular calibration of all the machinesand while cell to cell correlation testing is widely practised, the results have to be


judged on the basis of the degree of control over all critical variables external to theengine combustion chambers and by staff with practical experience in this type ofexercise.

Acceptance and type tests, power test codes and correctionfactors

Most of the work carried out in engine test installations is concerned with purelytechnical and engineering matters. Acceptance and type tests involve much more: theyare concerned with the interface between the engineering and commercial worlds andbetween manufacturer and customer. The reputation of the manufacturer is directlyinvolved and any ambiguity in the test procedures or in the interpretation of theresults can lead to the loss of goodwill and even to litigation. This is particularly thecase when the customer is a government department or one of the defence services.

It is therefore not surprising that tests of these kinds are the subject of extensiveregulations, in many cases having statutory force, and not always easy to interpret. Itis essential that the test facility manager who finds himself concerned with this kindof work should make himself thoroughly familiar with these regulations; the moreimportant English language documents are described briefly below.

Acceptance tests

These can range from the briefest acceptable running-in period plus a check on thegeneral operation of the engine, carried out in accordance with the engine manufac-turer’s internal procedures and without the involvement of the customer, to elaborateand detailed tests carried out in the presence of the customer or his representative inorder to ensure that the engine and its performance meet an agreed specification.

Type tests

Type tests are particularly associated with supplies to the Armed Forces, who invari-ably lay down the procedure in detail. A type test is designed to prove the wholeperformance of the engine under the conditions it may actually meet in service. Itshould cover everything necessary to prove beyond doubt or dispute that the requiredperformance is met or exceeded. This means that a type test will involve prolongedrunning under conditions representative of service with the specified maintenanceprocedures and measurements of wear and of such factors as piston cleanliness.

It will be quite distinct from the relatively brief production tests which, while alsochecking that the required performance is achieved, are aimed primarily at ensuringthat the quality of materials and workmanship and the production routine are up tothe standard established by the type test.

316 Engine Testing

Power test codes and correction factors

Several complex sets of rules are in general use for specifying the procedure formeasuring the performance of an engine and for correcting this to standard conditions.The most significant for the English-speaking world are:

BS 55141 Reciprocating internal combustion enginesISO 30462 Reciprocating internal combustion engines (identical to BS

5514)BS AU 1413 Road vehicle diesel enginesLloyd’s Rulebook4 (primarily concerned with marine engines and installations)

(American) SAE standards

SAE J19955 Engine power test code – spark ignition and compressionignition – gross power rating

SAE J13496 Engine power test code – spark ignition and compressionignition – net power rating.

BS 5514/ISO 3046

Perhaps the best starting point for the test engineer who finds himself involved intype testing or elaborate acceptance tests is a study of this standard (in either version).It is in six parts, briefly summarized below.

Part 1. Standard reference conditions, declarations of power, fuel andlubricating oil consumptions and test methods

In this part, standard atmospheric conditions are specified as follows:

• atmospheric pressure 1 bar (=750mmHg);• temperature 25�C (298K);• relative humidity 30 per cent.

The American SAE standards specify the same conditions (for marine engines stan-dard conditions are specified as 45�C (113�F) and 60 per cent relative humidity).

Specific fuel consumptions should be related to a lower calorific value of42 700 kJ/kg: alternatively the actual LCV should be quoted.

The various procedures for correcting power and fuel consumption to these con-ditions are laid down and a number of examples are given. These procedures areextremely complicated and not easy to use: they involve 45 different symbols, 17equations and 10 look-up tables. The procedures in the SAE standards listed aboveare only marginally less complex. Rigorous application of these rules involves agood deal of work and for everyday test work the ‘correction factors’ applied tomeasurements of combustion air consumption (the prime determinant of maximumengine power output) described in Chapter 12 will be found quite adequate.


Definitions are given for a number of different kinds of rated power: continuous,overload, service, ISO, etc., and a long list of auxiliaries which may or may notbe driven by the engine is provided. It is specified that in any declaration of brakepower the auxiliaries operating during the test should be listed and, in some cases,the power absorbed by the auxiliary should be given.

The section on test methods gives much detailed advice and instruction, includingtables listing measurements to be made, functional checks and tests for various specialpurposes. Finally, part 1 gives a useful check list of information to be supplied bythe customer and by the manufacturer, including the contents of the test report.

Part 3. Test measurements

This part discusses ‘accuracy’ in general terms, but consists largely in a tabulation ofall kinds of measurement associated with engine testing with, for each, a statement ofthe ‘permissible deviation’. The definition of this term is extremely limited: it definesthe range of values over which successive measurements made during a particulartest are allowed to vary for the test to be valid. These limits are by no means tightand are now outside the limits required of most R&D test facilities: e.g. ±3 per centfor power, ±3 per cent for specific fuel consumption; they thus have little relevanceto the general subject of accuracy.

Part 4. Speed governing

This part, which has been considerably elaborated in the 1997 issue, specifies fourlevels of governing accuracy, M1 to M4, and gives detailed instructions for carryingout the various tests.

Part 5. Torsional vibrations

This part is mainly concerned with defining the division of responsibility betweenthe engine manufacturer, the customer and the supplier of the ‘set’, or machinery tobe driven by the engine, e.g. a generator, compressor or ship propulsion system. Ingeneral, the supplier of the set is regarded as responsible for calculations and tests.

Part 6. Specification of overspeed protection

This part defines the various parameters associated with an overspeed protectionsystem. The requirements are to be agreed between engine manufacturer and customerand it is recommended that reset after overspeed should be manual.

Part 7. Codes for engine power

This part defines various letter codes, e.g. ICN for ISO standard power in English,French, Russian and German.

318 Engine Testing

It will be clear that this standard gives much valuable guidance regarding manyaspects of engine testing.

Statistical design of experiments

Throughout this book it has been the aim to give sufficient information concerningoften complex subjects to enable the reader to deal with the more straightforwardapplications without reference to other sources. In the case of the present topic thisis scarcely possible; all that can be done is to explain its significance and indicatewhere detailed guidance may be found.

The traditional rule of thumb in engineering development work has been to changeone thing at a time. It was thought that the best way to assess the effect of a designchange was to keep everything else fixed. There are disadvantages:

• It is very slow.• The information gained refers only to the one factor.• Even if we have hit on the optimum value it may change when other factors are

changed.

While many modern design of experiment (DoE) techniques and algorithms capableof optimizing multivariate problems are embedded in modern software suites that aredesigned to assist in engine mapping and calibration (see Chapter 19), it is consideredworthwhile describing one established method of approach.

Consider a typical multivariate development problem: optimizing the combustionsystem of a direct injection diesel engine, Fig. 15.3. The volume of the piston cavityis predetermined by the compression ratio but there is room for manoeuvre in:

Overall diameter D

Radius r

Depth d

d

h

φ

θ

r

D

Figure 15.3 Diesel engine combustion system


Angle of central cone �

Angle of fuel jets �

Height of injector h

Injection pressure p

not to mention other variables such as nozzle hole diameter and length and airswirl rate.

Clearly any attempt to optimize this design by changing one factor at a time islikely to over-run both the time and cost allocations and the question must be askedwhether there is some better way.

In fact, techniques for dealing with multivariate problems have been in existencefor a good many years. They were developed mainly in the field of agriculture,where the close control of experimental conditions is not possible and where asingle experiment can take a year. Published material from statisticians and biologistsnaturally tended to deal with examples from plant and animal breeding and did notappear to be immediately relevant to other disciplines.

The situation began to change in the 1980s, partly as a result of the work ofTaguchi, who applied these methods in the field of quality control. These ideas werepromoted by the American Supplier Institute and the term ‘Taguchi methods’ isoften applied loosely to any industrial experiment having a statistical basis. A clearexposition of the technique as applied to the sort of work with which the enginedeveloper is concerned is given by Grove and Davis7 who applied these methods inthe Ford Motor Company.

Returning to the case of the diesel combustion chamber, suppose that we aretaking a first look at the influence of the injector characteristics:

Included angle of fuel jets �

Height of injector h

Injection pressure p

For our initial tests we choose two values, denoted by + and −, for each factor, thevalues spanning our best guess at the optimum value. We then run a series of testsin accordance with Table 15.1. This is known as an orthogonal array, the feature ofwhich is that if we write +1 and −1 for each entry the sum of the products of anytwo columns, taken row by row, is zero.

The final column shows the result of the tests in terms of the dependent variable,the specific fuel consumption. The chosen values were:

�−= 110�

�+= 130�

h−= 2mmh+= 8mmp−= 800 barp+= 1200 bar

320 Engine Testing

Table 15.1 Test program for three parameters

Run � h p Specific

consumption

(g/kWh)

1 − − + 2072 + − + 2083 − + + 2104 + + + 2055 − − − 2186 + − − 2167 − + − 2208 + + − 212

One way of presenting these results is shown in Fig. 15.4 in which each of the fourpairs of results is plotted against each factor in turn. We can draw certain conclusionsfrom these plots. It is clear that injection pressure is the major influence on fuelconsumption and that there is some interaction between the three factors. There arevarious statistical procedures that can extract more information and guidance as tohow the experimental programme should continue.

The first step is to calculate the main effect of each factor. Line B of Fig. 15.4ashows that in this case the effect of changing the injection pressure from p− to p+is to reduce the specific consumption from 212 to 205 g/kWh. The main effect isdefined as one half this change or −3.5. The average main effect for the four pairsof tests is shown in Table 15.2.

A further characteristic of importance concerns the degree of interaction betweenthe various factors: the interaction between � and p is known as the �×p interaction.

Referring Fig. 15.4a, it is defined as one half the difference between the effect ofp on the specific consumption for �+ and the effect for �−, or

1/2��205−202�− �210−220��=+15

It would be zero if the lines in Fig. 15.4a were parallel. It may be shown that theinverse interaction p×� has the same value.

It may be shown that the sign of an interaction + or − is obtained by multiplyingtogether the signs of the two factors concerned. Table 15.2 shows the calculatedvalues of the three interactions in our example. It will be observed that the strongestinteraction, a negative one, is between h and �. This makes sense, since an increasein h coupled with a decrease in � will tend to direct the fuel jet to the same point inthe toroidal cavity of the combustion chamber.

This same technique for planning a series of tests may be applied to any numberof factors. Thus an orthogonal table may be constructed for all seven factors inFig. 15.3 and main effects and interactions calculated. The results will identify the


h–

h–

h–

h–

h+

h+

h+

h+

B

3

3

4

8

7Aφ–

φ+

φ+

φ+

φ+

φ+

φ+φ–

φ–

φ–

φ–

φ

φ–

p+

p+p+h+

p–h+

p–h–

p+h–

p–

p

h––

–– ++

+ +

p

(a)

220

200

200

g/k

Wh

g/k

Wh

220

(c) (d)

(b)

p–

77

3

31

1

1

4

4

2

2

2

88

6

6

6

55

5

Figure 15.4 Relationship between parameters showing main effects

Table 15.2 Main effects and interactions

Run � h p �×h �×p h× p

1 − − + + − −2 + − + − + −3 − + + − − +4 + + + + + +5 − − − + + +6 + − − − − +7 − + − − + −8 + + − + − −Main effect Interaction

−175 −025 −45 −15 +075 +025

322 Engine Testing

0

0

Figure 15.5 Engine map, torque, speed and fuel consumption

significant factors and indicate the direction in which to move for an optimum result.Tests run on these lines tend to yield far more information than simple ‘one variableat a time’ experiments. More advanced statistical analysis, beyond the scope of thepresent work, identifies which effects are genuine and which are the result of randomvariation.

A more elaborate version of the same method uses factors at three levels +, 0and −. This is a particularly valuable technique for such tasks as the mapping ofengine characteristics, since it permits the derivation of the coefficients of quadraticequations that describe the surface profile of the characteristic. Figure 15.5, repro-duced from Ref. 1, shows a part of an engine map relating speed, torque and fuelconsumption. The technique involves the choice of three values, equally spaced, ofeach factor. These are indicated in the figure.

It is hoped that the reader will gain some impression of the value of these methodsand gain some insight into the three-dimensional graphs that frequently appear inpapers relating to engine testing, from the above brief treatment.

Summary

The management structure of a test department is described and the planning, exe-cution and reporting of test programmes are discussed.

The correction of engine test results to international standards has been discussedand the statistical design of experiments is briefly introduced.


References

1. BS 5514 Parts 1 to 6 Reciprocating Internal Combustion Engines: Performance.

2. ISO 3046 Reciprocating Internal Combustion Engines: Performance.

3. BS AU 141a Specification for the Performance of Diesel Engines for Road Vehicles.

4. Rulebook, Chapter 8, Lloyd’s Register of Shipping, London.5. SAE Standard: Engine power test code – spark ignition and compression ignition – gross

power rating, SAE J1995 Jun 90.6. SAE Standard: Engine power test code – spark ignition and compression ignition – net

power rating, SAE J1349 Jun 90 (www.sae.org/certifiedpower).7. Grove, D.M. and Davis, T.P. (1992) Engineering Quality and Experimental Design,

Longmans, London.

16 Exhaust emissions

Introduction

It is probably true to say that, in the twenty-first century, a majority both of engineand vehicle development and of routine testing is concerned with environmentallegislation directed primarily towards the limitation and control of engine emissions.Testing may be divided between that directed towards development of engine andvehicle systems having the lowest possible environmental impact and that requiredby environmental legislation (homologation and certification) in order that a vehiclemay be sold in a particular world market.

Environmental legislation is particularly prolific in the field of vehicle emissions,but other areas, which include small two-stoke engines, medium-speed stationaryengines, marine engines and the engines of railway locomotives, are also receivingincreasing attention. The subject of engine and vehicle emissions is extremely widereaching and continues to generate a vast number of technical papers each year, asummary of which is well outside the scope of this book, but some are listed at theend of this chapter.

A gasoline-powered vehicle in 2005 was capable of emitting between 1 and 3per cent of the carbon monoxide (CO) of a comparable sized unit in 1970, similardramatic reductions are to be seen in vehicle emissions of hydrocarbons (HC), oxidesof nitrogen (NOx) and diesel particulates (all measured in mg/km).

Emission legislation in 2005 had divided into that developed in the United Statesand a generally more stringent form developed by the European Community (EU).Just over 50 per cent of the vehicles being sold in 2005 were being calibrated to EUemission standards. Asia-Pacific markets, with the exception of Taiwan, implementEU standards and some countries, such as Australia, currently allow both standards.

The recent history of such legislation has, naturally, been characterized by legis-lation setting future limits of exhaust emission components in advance of the engineand test technologies required to meet them, therefore it is impossible to summa-rize the list of components and systems that the test engineer will be using in thefuture. However, legislative approval of test technologies and techniques normallyruns somewhat slower than the development of instrumentation since the cost andorganizational complexity of developing and establishing a standard test procedure,even in a single country, are very significant; so once a procedure is established itis essential that it should remain unchanged for as long a period as possible. Thisprobably means current technology will continue to be used for some years but will

Exhaust emissions 325

mean that laboratories in future will be using the new technologies, such as massspectroscopy, for engine development while reserving other cell configurations forcertification.

The various harmful results of atmospheric pollution on the environment in gen-eral, and human health in particular, are now widely recorded and all readers of thisedition will be aware of them. The human health problems caused by polyaromatichydrocarbons (PAH) are a more recent discovery and have given rise to limits beingset in Europe (Directive 2003/17/EC) and various national regulatory bodies. The USEnvironmental Protection Agency (EPA) list of the PAH group includes 16 namedcompounds, the best known are benzene, formaldehyde and 1,3-butadiene, manyof which are known or suspected carcinogens. There has also been increased con-cern about the very small particulates, in the order of 0.1�m and smaller, so-called‘nanoparticles’ below 50 nm in size, that are able to penetrate to the surface of thelung (see Principles of particulate emissions measurement, below).

A reminder of the basic chemical transformation may be of interest to the reader.

Basic chemistry of internal combustion engine emissions

The two processes under review and shown in Fig. 16.1 are the complete and incom-plete combustion of the hydrocarbon fuel in air (treated here as an oxygen/nitrogenmixture). The production of carbon dioxide in this process is of concern since it isclassed as a ‘greenhouse’ gas, but can only be reduced by an increase in the overallefficiency of the engine and vehicle thereby producing lower fuel consumption perwork unit.

The emission gases and particulates covered by legislation are produced by incom-plete combustion and are

• carbon monoxide (CO), a highly toxic odourless gas;• carbon (C), experienced in the form of smoke;

Air

Fuel

Exhaust

Emissions

NOx

N

N

N

N

N

N

HH O

OC

C

C

C

OO O

OO

O

O

O O

H

NO2

NO

THCx y

CO

CO2

O2

N2

H2O

C (smoke)

Others

Complete

combustion

Incomplete

combustion

h

Hc

C

Figure 16.1 Gaseous components of combustion processes

326 Engine Testing

Accumulation

Nucleation mode Accumulation mode Heterogeneous particle

Diffusion

deposition

Adsorption

condensation

Figure 16.2 Phases of particle formation

• hydrocarbons (HC or TotalHC) formed by unburnt fractions of the originalliquid fuel;

• nitric oxide (NO) and nitrogen dioxide (NO2), together considered as NOx.

The physics and chemistry covering the development of particles between beingformed in the combustion chamber and circulated in the atmosphere is complex andhas to be emulated within the systems designed to measure exhaust emissions. Themain phases of particle formation are shown in Fig. 16.2; it is the products of theseprocesses that produce the gases and particles that may be absorbed by life forms‘on the street’.

Emissions from spark ignition engines

Perhaps the most characteristic feature of exhaust emissions, particularly in the caseof the spark ignition engine, is that almost every step that can be devised in order toreduce the amount of any given pollutant has undesirable side effects, most frequentlyan increase in some other pollutant.

Figure 16.3 shows the effect of changing air/fuel ratio on the emission of themain pollutants: CO, NOx and unburned hydrocarbons. It will be apparent that oneline of attack is to confine operation to a narrow window, around the stoichiometricratio, and a major thrust of vehicle engine development since the mid-1990s hasconcentrated on the so-called ‘stoichiometric’ engine, used in conjunction with thethree-way exhaust catalyser, which converts these pollutants to CO2 and nitrogen.

We have already seen in Chapter 14 that a spark ignition engine develops maxi-mum power with a rich mixture, � ∼ 0.9, and best economy with a weak mixture,� ∼ 1.1, and before exhaust pollution became a matter of concern it was the aim ofthe engine designer to run as close as possible to the latter condition except whenmaximum power was demanded. In both conditions the emissions of CO, NOx and


16

14

12

10

8

6

4

2

08

Water vapour

CO2

H2

NOx

O2

CO

Air/fuel ratio

Volu

me, per

cent of m

ois

t exhaust

Hydro

carb

ons a

nd N

Ox (

p.p

.m.)

Hydrocarbons

Sto

ichio

metr

ic

10 12 14 16 180

400

800

1200

1600

2000

2400

2800

Figure 16.3 Relation between exhaust emissions and air/fuel ratio for gasolineengines

unburned hydrocarbons are high: however, the catalyser requires precise control ofthe mixture strength to within about ±5 per cent of stoichiometric.

This requirement has been met by the introduction of fuel injection systems withelaborate arrangements to control mixture strength (lambda closed-loop control). Inaddition, detailed development of the inlet passages in the immediate neighbourhoodof the inlet valve and the adoption of four-valve heads has been aimed at improvingthe homogeneity of the mixture entering the cylinder and developing small-scaleturbulence: this improves the regularity of combustion, itself an important factor inreducing emissions. Other measures to improve the uniformity of the mixture includepreheating of the intake air and steps to reduce the extent of the liquid fuel film onthe intake passages.

Another area of development that significantly affects emissions concerns theignition system. Spark plug design and location, duration and energy of spark, aswell as the use of multiple spark plugs, all affect combustion and hence emissions,while spark timing has a powerful influence on fuel consumption, as well as emis-sions. Here again conflicting influences come into play. Delaying ignition after thatcorresponding to best efficiency reduces the production of NOx and unburned hydro-carbons, due to the continuation of combustion into the exhaust period, but increasesfuel consumption.

328 Engine Testing

Exhaust gas recirculation (EGR), involving the deflection of a proportion ofthe exhaust into the inlet manifold, reduces the NOx level, mainly as a result ofreduction in the maximum combustion temperature. The level of NOx production isvery sensitive to this temperature (this incidentally has discouraged development ofthe so-called adiabatic engine, in which fuel economy is improved by reducing in-cylinder heat losses). It also acts against the pursuit of higher efficiency by increasingcompression ratios.

EGR control calibration requires that the volume and content of the recirculatedportion of the exhaust flow ismeasured as a distinct channel for the gas analyser system.

An alternative line of development is concerned with the lean burn engine, whichoperates at lambda values of 1.4 or more, where NOx values have fallen to anacceptable value, and fuel consumption is acceptable. However, HC emissions arethen high, the power output per unit swept volume is reduced and running underlight load and idling conditions tends to be irregular. The engine depends for itsperformance on the development of a stratified charge, usually with in-cylinder fuelinjection.

Development of exhaust after-treatment systems is a continuing process, at presentparticularly concerned with reducing the time taken for the catalyser to reach operatingtemperature, important because emissions are particularly severe during cold starts.

Emissions from diesel engines

The composition of the emissions from a diesel engine, like that of the spark-ignitionengine, depends on the engine design, its operating condition and the compositionof the fuel used, the latter property being particularly important since it is veryvariable worldwide. The sulphur content of diesel fuel is responsible for sulphurdioxide, SO2, in the exhaust. Permitted levels of sulphur in fuels for road vehicleshave been drastically reduced and have given rise to incidental problems with fuelinjection equipment arising from the reduced lubricity of the fuel that has been, andcontinues to be, the subject of engine testing in cells and in real-life operation. Inthe developing world, sulphur contents tend to be much higher and an obstacle toreduction is the substantial cost of the necessary refinery modifications.

The diesel engine presents rather different problems from the spark-ignition enginebecause it always operates with considerable excess air, so that CO emissions arenot a significant problem, and the close control of air/fuel ratio, so significant in thecontrol of gasoline engine emissions, is not required. On the other hand, particulatesare much more of a problem and NOx production is substantial.

The indirect injection (prechamber) engine performs well in terms of NOx emis-sions; however, the fuel consumption penalty associated with indirect injection hasresulted in a general move to direct injection, associated with four-valve cylinderheads with a central fuel injector. This development is associated with a sharp increasein fuel injection pressures, now commonly 1500 bar or more, and the ‘shaping’ of thefuel injection pulse. The ‘common rail’ fuel injection system, where fuel is supplied


at a uniform high pressure and injector needle movement controlled electrically, isnow the most common fuelling system in small automotive applications.

NOx emissions are very sensitive to maximum cylinder temperature and to theexcess air factor. This has prompted the use of increased levels of turbochargingwith improved after-cooling, as well as the use of retarded injection which resultsin reduced peak pressures and temperatures but, beyond a certain point, in increasedfuel consumption.

Diesel particulate emissions

Individual diesel exhaust particles may be highly complex coagulations of com-pounds, as shown in Fig. 16.4 below. Around 90 per cent of particles emitted bya modern automotive diesel may be below 1�m in size which challenges the toolsused to measure their presence.

Emission legislation, certification and test processes

This is an immensely complicated and rapidly evolving field: all but the briefestsummary would be impossible within the limits of this book, and would in any casebe out of date almost as soon as it was published. The reader requiring detailedknowledge is advised to consult Ref. 1. This publication summarizes internationalstandards originating from the USA, the UN Economic Commission for Europe(ECE) and the European Union (EU), together with national standards for some 18individual countries.

Since so much of the operational strategy of automotive test facilities is determinedby current and planned environmental protection legislation, it is necessary for test

Fractal-likesoot agglomerate

50–500 nm

Liquid hydrocarbonparticle 10–80 nm

(–5 nm soot or ash core)1. attached to soot2. freely in air

1. attached to soot2. freely in air

Adsorbed hydrocarbons

Liquid sulphate particle

Carbon spheres(including metals)

10–80 nm

Figure 16.4 Schematic representation of diesel particles (Kulijk–Foster)

330 Engine Testing

engineers to be aware of the legislative framework under which their home and exportmarkets work. A single source of up-to-date information on worldwide emissionlegislation may not exist in the public domain, therefore the industry has to supportspecialists in particular areas of commercial interest in order to keep up to date withingeographical zones.

Vehicle emission legislation dates from 1966 when the California Air ResourcesBoard (CARB) produced tailpipe emission standards for hydrocarbons (HC) andcarbon monoxide (CO) within the State of California. The Environmental ProtectionAgency (EPA) and the Clean Air Act were introduced by the US Government in the1960s. European and Japanese legislation followed from 1970. The increasing publicawareness of the environmental and human health problems resulting from vehicularemissions plus legislative barriers to important automotive markets began a devel-opment impetus that produced, and continues to produce, new emission-reducingtechnologies and requiring improved fuels. Catalytic converters required lead-freegasoline, thus the whole support infrastructure and pattern of vehicle registrationhas increasingly become dominated by emission legislation; the process wherebyautomotive technology, test technology and legislative requirements ‘leapfrog’ eachother continues to this day.

The huge improvements in individual engine emissions between 1965 and 2005have meant that now the refining and formulation of lubricating oils and fuels haveto be improved worldwide in order to benefit from the modern combustion controland after-treatment technologies.

In spite of improved instrumentation and much reduced levels of pollutants pro-duced by vehicles and engines, the basic form of much legislative testing has remainedvery similar to the original EPA methodology of the 1970s.

In general, all engine emission legislation consists of three component parts(Fig. 16.5):

1. Test cycle describes the operation of the tested vehicle or engine. For light dutyvehicles, it simulates the actual driving on the road in that it defines a vehicle

Figure 16.5 The three pillars of emission legislation


velocity profile over the test time. For heavy duty and off-road engines whereonly the engine is tested on an engine dynamometer, the test cycle defines aspeed and torque profile over the test time.

2. Test procedure defines in detail how the test is executed, which measurementmethod and which test systems have to be used. It defines the test conditions andresult calculations to apply. Probably the best known example of a legislativeprocedure is the use of sample bags in which an accumulation of diluted sampleof exhaust gas resulting from a drive cycle is stored for analysis of content andconcentration.

3. Test limits, which define the maximum allowed emission of the regulated com-ponents in the engine exhaust. For light duty vehicles, the limit is expressedin mass per driving distance (g/km); for heavy duty vehicles, the limits areexpressed in mass per unit of work (g/kWh).

Light duty certification requires the whole vehicle to be tested and certified, whileheavy duty engines are tested in a test cell operated in accordance with the stipulatedtest procedures.

The basic legislation methodology for cars and light truck engines was, and stillis, based on driving the test vehicle through a prescribed ‘drive cycle’ on a chassisdynamometer (see Chapter 18), while storing a proportion of the exhaust gases,produced during the whole test, in bags so that the amount and proportions of thekey pollutants can be determined.

The drive cycles were designed to simulate typical vehicle usage, but since theyhad to be driven as consistently as possible over a wide variety of vehicles, the rateof change in simulated road speeds is quite slow when compared with real life. Ithas only been since the beginning of the twenty-first century that instrumentationhas been available that is truly capable of measuring some of the exhaust pollutantsas they are produced during the transient phases of engine loads.

Human drivers using a drivers-aid display rather like an early generation computergame display are still employed, but increasingly test vehicles are driven usingelectromechanical robots.

Drive cycles have been subject to much international examination by nationallaboratories around the world, it being obvious that a ‘typical’ drive for a suburbancommuter living in Los Angeles will be quite different to that of an inhabitant ofMumbai or Cairns; however, the majority of testing worldwide is done to drive cyclesproduced by the three dominating geographical producer areas.

It was understood from the early days of emission studies that some of thechemical reactions resulting from in-cylinder combustion continue, as the gases andparticularly ‘soot’ particles, pass through the exhaust system and into the atmosphere.Therefore, the pollution products had to be subjected to realistic post-combustionmixing with air before being analysed. The complex chemistry and the need forcorrelation across many test sites worldwide has meant that much of the legislationconcerning vehicle homologation is very prescriptive concerning equipment used.There is, therefore, a lag between the instrumentation developed for emission researchand that ‘allowed’ for certification by legislation.

332 Engine Testing

Legislation classifications

Much of the test subject classification in automotive emission legislation is by vehiclesize rather than engine size and type. The main classifications being

• light duty – gasoline;• light duty – diesel;• heavy duty.

Each of these categories has legislative pollutant limits, test methodologies andinstrumentation designed specifically for them.

In European emission legislation, there are both categories and weight classifica-tions of vehicles which are shown below:

• Category M: Motor vehicles with at least four wheels and used for the carriageof passengers.

• Category M1: Vehicles used for the carriage of passengers and comprising nomore than eight seats in addition to the driver’s seat.

• Category M2: Vehicles used for the carriage of passengers and comprising nomore than eight seats in addition to the driver’s seat and a maximum mass notexceeding 5 tonnes.

• Category M3: Vehicles used for the carriage of passengers and comprising nomore than eight seats in addition to the driver’s seat and a maximum massexceeding 5 tonnes.

• Category N: Motor vehicles with at least four wheels and used for the carriageof goods.

• Category N1: Vehicles used for the carriage of goods having a maximum massnot exceeding 3.5 tonnes.

• Category N2: Vehicles used for the carriage of goods having a maximum massexceeding 3.5 tonnes and not exceeding 12 tonnes.

• Category N3: Vehicles used for the carriage of goods having a maximum massexceeding 12 tonnes.

Weight classifications for light commercial vehicles:

• Class I RW ≤1305 kg• Class II 1305 < RW ≤ 1760 kg• Class III 1760 kg < RW

Tables 16.1a, b and c list some examples of emission limits for conventional vehi-cles classified above. Note that the dates refer to new vehicle types; new vehicleregistrations are 1 year later. Euro III and IV standards also apply to Class I lightcommercial vehicles <1305 kg.


The classification of vehicles meeting various emission limits has given rise to anumber of acronyms that are commonly used in technical press and papers:

Low emission vehicle LEVUltra low emission vehicle ULEVSuper ultra low emission vehicle SULEVTransitional low emission vehicle TLEVZero emission vehicle ZEVPartial zero emission vehicle PZEV

Table 16.1a European emission limits for petrol cars in grams per kilometre

Gasoline As from CO THC NOx

Euro I 1/7/1992 4.05 0.66 0.49Euro II 1/1/1996 3.28 0.34 0.25Euro III 1/1/2000 2.30 0.20 0.15Euro IV 1/1/2005 1.00 0.10 0.08

Table 16.1b European emission limits for diesel cars in grams per kilometre

Diesel As from CO THC NOx PM

Euro I 1/7/1992 2.88 0.20 0.78 0�14Euro II 1/1/1996 1.06 0.19 0.73 0�10Euro III 1/1/2000 0.64 0.06 0.50 0�05Euro IV 1/1/2005 0.50 0.05 0.25 0�025

Table 16.1c European emission limits for light commercial vehicles N2 Classin grams per kilometre

N2 As from Fuel type CO HC NOx HC+NOx PM

Euro I 1/10/1994 All 5�17 − − 1.4 0.19

Euro II 1/1/1998Petrol 4 − − 0.65 −Diesel 1�2 − − 1.1 0.15

Euro III 1/1/2002Petrol 4�17 0.25 0.18 − −Diesel 0�8 − 0.65 0.72 0.07

Euro IV 1/1/2006Petrol 1�81 0.13 0.1 − −Diesel 0�63 − 0.33 0.39 0.04

334 Engine Testing

In the 1970s the net was cast further and legislation covering small utility engines,being spark-ignition engines, both two- and four-stroke, of the type used in mowers,chainsaws and small generators was produced and continues to be developed. InEurope, small engines under 19 kW were covered in 2002 (Directive 2002/88/EC).

Legislation covering ‘non-road’ diesel engines was produced in Europe in 1997(Directive 97/68/EC) and intended to cover off-road trucks, drilling rigs, mobilecranes, etc. The coverage is complex in detail and much is phased in over time toallow for replacement engines and specialist usages, but much of the test detailsare based on ISO 8178-1 which is an international standard for non-road engineapplications, specifically:

• EPA CFR 40 Part 89 Non-road compression ignition engines;• EPA CFR 40 Part 90 Non-road spark ignition engines;• ECE 97/68 amended by ECE 202/88.

The engine test sequences based on ISO 8178 are sequences of steady state modeswith different emission limits. A typical set-up for running these tests in a cell isshown in Fig. 16.6.

Engineers requiring details need access to the relevant government websites cov-ering local legislation.

International agreement is clearly required for the control of marine plant sinceit can travel the world; such agreement is contained in the 1997 MARPOL ProtocolAnnex V1. Marine engine legislation is divided into three categories defined bysize (displacement per cylinder) but covering the three distinctly different enginetechnologies (Table 16.2).

Category 3 has caused difficulty between environmental groups and legislatorssince it covers very large diesels running on residual fuels (see Chapter 7), theformulation of which is largely unregulated. Many countries apply rules based on

Dynamometer

CLDlc

CO2 high

COlow

FIDlc

AVL

Air intake flowmeasurement

Liquid fuel flowmeasurement

Opacimeter

Partial flowparticulatesampling

Emissionbench

Figure 16.6 Typical emission plant used in ISO 8178 engine emission cycles(diesel)


Table 16.2 Marine engine categories

Category Displacement per cylinder (D) Basic engine technology

1 D <5 dm3, power ≥37 kW Land-based non-road diesel2 D ≥5 dm3 and <30 dm3 Locomotive diesel engine3 D ≥30 dm3 Unique marine engine design

ISO 8178 for ships within coastal waters and habour which requires ships runningon residual fuels to change over to light distillate oils before entering waters coveredby such legislation.

Aftermarket emission legislation

In addition to the prime legislation covering the certification of new engines andvehicle systems in most countries of the developed world, there is emission legislationcovering the condition of cars in the population as they age. Such tests are requiredannually and range from a single visual smoke check, through a check at fast idlefor levels of CO and HC, to a test under light load on a rolling road that checks CO,CO2 and HC. In the UK, the basic emission test (BET) for gasoline engines is asfollows:

Engine warmed upFast idle 2500–3000 r.p.m., CO no more than 0.3 per cent, HC no more than 200

p.p.m.Lambda between 0.97 and 1.03Normal idle 450–1500 r.p.m., CO no more than 0.5 per cent.

For diesel engines, the test checks for visible smoke during idle and transient statesby using a calibrated smoke meter placed in the tailpipe, while the engine is subjectedto a number of accelerations.

Homologation

Homologation is a term widely thought to be exclusively involved with certifica-tion to emission legislation; in fact it is another case of a word, having originallyno particular engineering associations, being taken over and given a specializedmeaning.

Knowledge of and compliance with the legislative requirements of different mar-kets, covering exhaust emissions but also safety, fuel consumption rates, noise vibra-tion and harshness, competitive benchmarking, etc., is an important aspect of theproblem of gaining acceptance of a given product. For the engine or vehicle manu-facturer, homologation is a complex and expensive area of activity since all major

336 Engine Testing

versions, and all derivatives, of the vehicle must meet the formal requirements thatare in force in each country in which the vehicle is to be sold.

Homologation is the process of establishing and certifying this conformity, bothfor whole vehicles and for components.

Principles of particulate emissions measurement

Particulates, when they appear to the human observer, are called ‘smoke’. Smokecolours are indicative of the dominant source of particulate:

• black= ‘soot’ or more accurately carbon, which typically makes up some 95 percent of diesel smoke either in elemental, the majority, or organic form;

• blue= hydrocarbons, typically due to lubricating oil burning due to an enginefault;

• white=water vapour, typically from condensation in a cold engine or coolantleaking into the combustion chambers – white smoke is not detected by conven-tional smoke meters;

• brown=NO2 may be detected in exhaust of heavy fuel engines.

There are essentially four methods in use for measuring particulate emissions andnumerical results produced by these methods cannot readily be related one to theother:

• Opacimeters that measure the opacity of the undiluted exhaust by the degree ofobscuration of a light beam. These devices are able to detect particulate levelsin gas flow at lower levels than the human eye can detect. The value output isnormally in percentage of light blocked by the test flow. Zero being clean purgeair and 100 per cent being very thick black smoke. A secondary output giving anabsorption factor ‘k’ m−1 may be given, which allows some degree of comparisonbetween devices since it removes the effect of different distances between lightsource and sensor.

• Smoke meters that perform the measurement of the particulate content of anundiluted sample of exhaust gas by drawing it through a filter paper of specifiedproperties and estimating the consequent blackening of the paper against a pristinepaper. The value output is in some form of ‘smoke number’ specific to theinstrument maker.

• Particulate samplers that measure the actual mass of particulates trapped by afilter paper during the passage of a specified volume of diluted exhaust gas. Thisrequires climatically controlled laboratory weighing facilities.

• The ever lower particulate emissions from engines has required a new generationof ‘microsoot’ devices capable of detecting particulate levels down to typically5�g per m3 of exhaust gas. These devices incorporate a laser and work on aphoto-acoustic principle.


Principles of measurement and analysis of gaseous emissions

The description of instrumentation that follows, which is not exhaustive, lists themain types of instrument used for measuring the various gaseous components ofexhaust emissions. Continuing developments are taking place in this field, notablyin the evolution of instruments having the shortest possible response time.

Non-dispersive infrared analyser (NDIR)

Called ‘non-dispersive’ because all the polychromatic light from the source passesthrough the gas sample before going through a filter in front of the sensor, whereas‘dispersive’ instruments, found in analytical laboratories, filter the source light to anarrow frequency band before the sample.

The CO2 molecule has a very marked and unique absorptance band of infrared(IR) light that shows a dominant peak at the 4.26�m wavelength which the instru-ment sensor is tuned to detect and measure. By selecting filters sensitive to otherwavelengths of IR, it is possible to detect other compounds such as CO and otherhydrocarbons at around 3.4�m (see Fourier transform infrared analyser, below).Note that the measurement of CO2 using an NDIR analyser is cross-sensitive to thepresence of water vapour in the sample gas.

Fourier transform infrared analyser (FTIR)

This operates on the same principle as the NDIR, but performs a Fourier analysisof the complete infrared absorption spectrum of the gas sample. This permits themeasurement of the content of a large number of different components. The methodis particularly useful for dealing with emissions from engines burning alcohol-basedfuels, since methanol and formaldehyde may be detected.

Chemiluminescence detector (CLD)

Chemiluminescence is the phenomenon by which some chemical reactions producelight. The reaction of interest to exhaust emissions is

NO+O3 → NO2+O2 → NO2+O2+photon

The nitrogen compounds in exhaust gas are a mixture of NO and NO2, describedas NOx.

In the detector, the NO2 is first catalytically converted to NO and the sample isreacted with ozone which is generated by an electrical discharge through oxygen, atlow pressure in a heated vacuum chamber. The light is measured by a photomultiplierand indicates the NOx concentration in the sample.

338 Engine Testing

A great deal of development work continues to be carried out to improve chemicalreaction times which are highly temperature-dependent, and so shorten instrumentresponse times.

Flame ionization detector (FID)

The FID has a very wide dynamic range and high sensitivity to all substances thatcontain carbon. The operation of this instrument shown schematically in Fig. 16.7depends on the production of free electrons and positive ions that takes place duringthe combustion of hydrocarbons. If the combustion is arranged to take place in anelectric field, the current flow between anode and cathode is closely proportional tothe number of carbon atoms taking part in the reaction. In the detector the sampleis mixed with hydrogen and helium and burned in a chamber which is heated toprevent condensation of the water vapour formed. A typical, sample to measurement,response time is 1–2 s.

Fast FID, cutter FID and GC-FID

This is a miniaturized development of the FID instrument which is capable of aresponse time measured in milliseconds and may thus be used for in-cylinder and

Thermocouple

Ignition

Amplifier

HT power supply

Air

H2

Sample gas

Burner

Collector

electrode

Figure 16.7 Flame ionization detector


exhaust port measurements. Cutter FID analysers and GC-FID analysers are used tomeasure methane CH4. In the fast FID, non-methane hydrocarbons are catalysed intoCH4 or CO2 and measured.

Paramagnetic detection (PMD) analyser

PMD analysers are used to measure oxygen in the testing of gasoline engines. Theywork due to the fact that oxygen has a strong paramagnetic susceptibility. Insidethe measuring cell, the oxygen molecules are drawn into a strong inhomogeneousmagnetic field where they tend to collect in the area of strongest flux and physicallydisplace a balanced detector whose deflection is proportional to oxygen concentration.Since NOx and CO2 show some paramagnetic characteristic, the analyser has to becapable of calculating compensation for this interference.

Mass spectrometer

These devices, not yet specified in any emission legislation, are developing rapidlyand can distinguish most of the components of automotive engine exhaust gases;currently they remain an R&D tool, but may represent the future technology ofgeneral emission measurement.

Response times

All these instruments are required to operate under one of two quite different condi-tions, depending on whether the demand is for an accurate measurement of a samplecollected over a fairly long time interval, as in the case of the various statutorytest procedures, or for an instantaneous measurement made under rapidly changingconditions, such as arise in true transient testing, the purpose of which is to studythe performance of the engine and its control system in detail.

These two sets of requirements are conflicting: analysers for steady state workmust be accurate, sensitive and stable and thus tend to have slow response times andbe well damped. Analysers for transient work do not require such a high standardof accuracy but must respond very quickly, preferably within a few milliseconds.Reduction of response time is a prime object of instrument development in thisfield.

A further problem in transient testing is concerned with the time and distance lagsassociated with the positioning of the exhaust gas sampling points. As any search willreveal, this matter is discussed in many SAE papers concerning the reconstructionof the true signal from the instrument signals, taking into account sampling delaysand instrument response characteristics.

340 Engine Testing

Integration of exhaust emission instrumentation

The great majority of engine test work is concerned with the taking of measurementsthat are in principle quite simple, even though great skill may be needed to interpretthem: forces, pressures, masses, flow rates, speeds, displacements, temperatures,oscillations. Where emissions measurements are concerned, we are forced to moveinto a totally different and very sophisticated field of instrumentation engineering.The apparatus makes use of subtle and difficult techniques borrowed from the fieldof physics.

It is very desirable, for all but the simplest garage-type emissions apparatus, thata technician should be specially trained to take responsibility for the maintenanceand calibration of the instruments involved.

There are significant subsystems that require to be housed in engine or vehicletest facilities, each of which require some special health and safety considerationsdue to the handling of hot and toxic gases within a building space. These will beconsidered in the following text.

The full flow of exhaust gases has to be safely ducted and samples for analysishave to be removed in the correct amounts at critical points in any vehicular after-treatment system.

It is important that water, which is a product of the combustion process, particularlythat of the gasoline and some biofuels, is not allowed to condense within the samplingsystem. Some of the compounds produced by the partial combustion of fuels aresoluble in water, so it is possible to lose a significant amount of these compounds inthe condensation on the equipment surfaces.

To prevent condensation in critical parts of the exhaust, the sampling system willbe heated. For example, the sampling system for THC may consist of heated probe,sampling line filter and sampling pump, all of which are required to prevent thetemperature of the equipment being lower than the dew point of the exhaust gas.

Calibration, span gases, storing gas distribution system

In order to function correctly exhaust gas analysers have to be calibrated regularly,using gases of known composition. The principle of calibration is similar to thatof any transducer in that the zero point is set, in this case by purging with purenitrogen (sometimes referred to as ‘zero-air’), then the 100 per cent value is set witha gas of that composition, known as a ‘span gas’. Good practice then dictates thata gas of an intermediate composition is then used as a check of system linearity.The whole routine may be highly automated by a calibration routine built into theanalyser control unit. After calibration, the gas supply system is purged of at leastNOx to prevent degrading in the lines.

Many R&D analysers are capable of part-spanning to change the measuringsensitivity and can be calibrated by using either known span gas precisely dilutedwith a zero-air using a device known as a ‘gas divider’ or, more commonly, with


Table 16.3 List of typical gases that may be used in the operation of basicengine emission certification cell

Gases Used as Concentration For analyser type

Synthetic air (2) Burner air 21% O2/79% N2 FIDH2/He mixture Burner gas 40% H2, 60% He FIDC3H8 in synth� air Test gas 99.95% External bottle with

CFO systemO2 Operating gas 99.98% purity CLDCO in N2 Span gas 50 ppm COlow

C3H8 in synth air Span gas 100 ppm FIDC3H8 in synth air Span gas 10000 ppm FIDNO in N2 Span gas 100 ppm CLDNO in N2 Span gas 5000 ppm CLDCO in N2 Span gas 500 ppm COlow

CO in N2 Span gas 10% COhigh

CO2 in N2 Span gas 5% CO2

CO2 in N2 Span gas 20% CO2

O2 in N2 Span gas 20% CO2

Synthetic air (1) Zero gas oil free FIDN2 Zero gas CO free NO, COlow, COhigh,

CO2, O2

a bottled span gas of the correct composition. If using the gas divider method, it isimportant that the same gas (N2 or artificial air) is used in the dilution as that whichwas used as the zero gas.

In addition to calibration or span gases, the facility will require operational gasessuch as hydrogen plus an oxygen source as fuel for any flame ionization (FID) andin some cases test gases for such routines as a critical flow check (see ConstantVolume Sampling (CVS) systems).

Table 16.3 shows a possible minimum list of gases required in a modern emissionstest facility.

All these gases have to be stored at high pressure, usually between 150 and 200bar, and distributed, at a regulated pressure of about 4 bar, to the analysis deviceswithin the facility. The pipe work system used in the distribution has to be chemicallyinert and medically clean to prevent compromising the composition of the calibrationgases. The minimum number of 50 litre or 40 litre

∗gas storage bottles containing

individual gases is probably 12, and a typical diesel/gasoline certification cell mayhave upwards of 20 gas lines feeding its analyser system. It is not uncommon in

∗A 40 litre bottle operating at 150 bar has a capacity of around 6000 litres of gas at theanalyser pressure and a 50 litre bottle (normal size for synthetic air and nitrogen) around7500 litres.

342 Engine Testing

mixed fuel facilities with two cells to have 20 calibration bottles and four operatinggases connected at any one time.

Each gas bottle has to be delivered, connected, disconnected and removed in thecycle of use; therefore the positioning of the gas store has some important logisticalconsiderations.

Clearly, the distance between the gas bottles, the pressure regulation station andanalysers directly affects the cost of gas distribution. The practical requirements ofrouting upwards of twelve 1

4 -inch or 6-mm OD stainless steel pipes on a supportframe system is an important design consideration in a new facility. The lengthof gas lines may be sufficient to create an unacceptable pressure drop which mayrequire transmission at an intermediate pressure and regulation local to the analysers.Clearly, the location of the gas bottle store or stores can have important cost andoperational implications on the facility.

The storage of the gas bottles should be in a well-ventilated secure store wherethe bottles are not subjected to direct solar heating and thus remain under 50�C.For operational reasons, in very cold climates the bottle store needs to be kept attemperatures that allow handling and operation without injury to the operator orregulation malfunction. The store commonly takes the form of one or more roofed‘lean-to’ structures on the outside wall of the facility, in temperate climates havingsteel mesh walls and always positioned for easy access of delivery trucks. Storageof gas bottles above ground within buildings creates operational and safety problemsthat are best avoided.

Pressure regulators may be either wall mounted with a short length of high pressure(up to 200 bar) flexible hose connecting the cylinder or, more rarely, directly cylindermounted. The regulators handling NOx gases need to be constructed of stainless steel,while other gases can be handled with brass and steel regulators. Local regulationsand planning permission requirements may apply.

The gas lines are usually made of stainless steel specifically made for the purpose;those transporting NOx, sometimes all, are electropolished internally to minimizedegradation and reaction between gas and pipe. Pipes have to be joined by orbitalwelding

∗∗or fittings made expressly for the purpose. The pipes should terminate

near the emission bench position in a manifold of isolating valves and self-sealingconnectors for the short flexible connector lines to the analyser made of inert Teflonor PTFE not more than 2m long.

Gas analysers and emission benches

Although often used interchangeably, the term ‘analyser’ should refer to an individualpiece of instrumentation dedicated to a single analytical task installed within an‘emission bench’ that contains several such units. Most analysers fit within a standard

∗∗Orbital welding uses a gas tungsten arc welding (GTAW) process that prevents oxidationand minimizes surface distortion of pipe interior.


Table 16.4 Typical analyser measuring minimum and maximum ranges

Analyser Low range High rangeNDIR COlow 0–100 p.p.m. 0–3000 p.p.m.

COhigh 0–1% 0–12%CO2 0–2% 0–20%NO 0–100 p.p.m. 0–5000 p.p.m.

PMD O2 0–10% 0–25%FID heated THC 0–100 p.p.m. 0–10 000 p.p.m.CLD NOx 0–100 p.p.m. 0–5000 p.p.m.

19-inch rack and are typically 4 HU high; most are capable of being set for measuringdifferent ranges of their specific chemical. The Table 16.4 shows what analysersmight be installed within a high quality bench built to measure raw gas from enginestested under ISO 8178 cycles.

Constant volume sampling (CVS) systems

Clearly, the above methods give no indication, except by inference based on experi-ence, of the actual mass of particulates present in a given volume of exhaust. In bothdiesel and gasoline testing we need to dilute the exhaust with ambient air to preventcondensation of water in the collection system. It is necessary to measure or controlthe total volume of exhaust plus dilution air and collect a continuously proportionedvolume of the sample for analysis. The use of a full flow CVS system is mandatoryin some legislation, particularly that produced by the EPA. It may be assumed thatthis will change over time with developing technology and ‘mini-dilution systems’will become widely allowed.

CVS systems consist of the following major component parts:

• A tunnel inlet air filter in the case of diesel testing or a filter/mixing tee in thecase of gasoline testing, which mixes the exhaust gas and dilution air in a ratiousually about 4:1. There is also a sample point to draw off some of the (ambient)dilution air for later analysis from a sample bag.

• The dilution tunnel, made of polished stainless steel and of sufficient size toencourage thorough mixing and to reduce the sample temperature to about 125�F(51.7�C). It is important to prevent condensation of water in the tunnel so in thecase of systems designed for use in climatic cells, the air will be heated beforemixing with gas.

• A proportion of the diluted exhaust is extracted by the bag sampling unit forstorage in sample bags.

• The critical flow venturi controls and measures the flow of gas that the turbo-blower draws through the system.

344 Engine Testing

• The gas sample storage bag array.• The analyser and control system by which the mass of HC, CO, NOx, CO2 and

CH4 is calculated from gas concentration in the bags, the gas density and totalvolume, taking into account the composition of the dilution air component.

• In the case of diesel testing, part of the flow is taken off to the dilution samplercontaining filter papers for determining the mass of particulates over the testcycle.

A typical layout of a CVS system based on a 2× 2 chassis dynamometer is shownin Fig. 16.8. The arrangement shown is for a cell certifying vehicles, either diesel orgasoline, to Euro 2 or Euro 3 regulation. The system shown includes critical flowventuri (marked CFV) through which the gas/air is sucked by a fixed speed ‘blower’fitted with an outlet silencer. Some systems control flow with a variable speed blowerand a different type of ‘subsonic’ venturi.

US Federal regulations require that the flow and dilution rates can be testedby incorporation of a critical flow orifice (CFO) check. This requires a subsidiarystainless steel circuit that precisely injects 99.95 per cent propane as a test gasupstream of the mixing point, the CVS system dilutes this gas according to the flowrate setting allowing the performance to be checked at a stabilized temperature.

The full flow particulate tunnel for heavy vehicle engines is a very bulky device.The modules listed above may be dispersed within a test facility because of theconstraints of the building. The tunnel has to be in the cell near the engine andthere are some legislative requirements concerning the distance from the tail pipeand the dilution point, ranging from 6.1m for light duty to about 10m for heavyduty systems.

Similarly, there are limits in some legislation concerning the transit time for HCsamples from the sample probe to the analyser of 4 s, which will determine theposition within the building spaces of the FID unit.

The analysers may be in the control room, the venturi and sample bags withinanother room adjoining the control room. The turboblower is commonly placed onthe roof since they often produce high noise levels. Alternatively, for light vehicleengines the analysers, CVS control and bag storage may be packaged in one, typicallyfour-bay, instrument cabinet.

The so-called ‘mini-dilution tunnel’ has been developed to reduce the problemsinherent in CVS systems within existing test facilities. However, users must checkif its use is authorized by the legislation to which they are working.

It is clearly necessary in this case to measure precisely the proportion of the totalexhaust flow entering the tunnel. This is achieved by the use of a specially designedsampling probe system that ensures that the ratio between the sampling flow rateand the exhaust gas flow rate is kept constant

In all CVS systems, exhaust gases may be sent to the analysers for measurementeither from the accumulated volume in the sample bags or directly from the vehicle,so-called ‘modal’ sampling.

Chassis Dyno Controller

Rack -

Mount PC -

500

Rack -

Mount

PC -500

Device controller

USV

Driver’s aid

Test controller TCP / IP

VGA

RS 232

RS 232

RS 232An

alo

gu

e 0

-10

V

Silencer

Heated THC measurement

Weather station

3 Ambient Air Bags

3Exhaust

Bags

Printer

TCP / IP

Exhaust bagsAmbient bags

Exhaust modal

CLDlc

CO2 high

CO low

FIDlc

Gasoline

Diesel

Dilution tunnel 273

Blower

Modal

AmbientBags

Driver’s aid

display

FID

Dynamometer controller

Mixing tee

Selectable CF venturi

Temperature control heat exchanger

3 x

Particulate

sampling

Heated THC probe

Cooling fan

RS 232

RS 232

3

Figure 16.8 The major components of a Euro 3 CVS system

346 Engine Testing

In an advanced test facility such as those involved with SULEV development,there will be three to five tapping points created in the vehicle system, from whichgas may be drawn for analysis such as

1. exhaust gas recirculation sample (EGR);2. before the vehicle catalytic converter (pre-CAT);3. after the vehicle catalytic converter (post-CAT);4. tail pipe sample (modal);5. diluted sample (sample bags).

From 2 and 3 the efficiency of the catalyst may be calculated.In addition to the tapping points listed above, there may be several more taking

samples from the CVS tunnel for particulate measurement.

Health and safety implications and interlocks for CVS systems

The following three fault conditions are drawn from experience and should beincluded in any operational risk analysis of a new facility.

Where the exhaust dilution air is drawn from the test cell and in the fault conditionof the blower working and the cell ventilation system intake being out of action, orswitched off, a negative cell pressure can be created that will prevent the openingof cell doors. The cell ventilation system should therefore be interlocked with theblower so as to prevent this occurring.

In the fault condition of the blower being inoperative or the airflow too low due totoo small a flow venturi being used, the hot exhaust gas can flow back into the filterand cause a fire. A temperature transducer should be fitted immediately downstreamof the dilution tunnel inlet filter and connected to an alarm channel set at about 45�C.

Gas analysers that have to discharge their calibration gases at atmospheric pressureshould have an extraction cowl over them to duct the gases into the ventilationextract; such an extract should be interlocked with the analyser, preferably by a flowsensor.

Emissions test procedures

The distinguishing feature of all emissions test procedures is that they all involvetransient operation of the engine. They are thus more demanding than any sort ofsteady state test. They are, however, representative of the conditions of operation ofvehicle engines.

A comprehensive survey of current and proposed test procedures would muchexceed the scope of this book. It is therefore proposed to describe the two principalcurrent test procedures for passenger cars and to give a brief indication of the scopeof other methods and anticipated developments.


Temperature soak areas for legislative testing

A feature of many legislative test procedures is that the vehicle has to soak at a uniformtemperature before commencement. The two temperatures stipulated are between 20and30�Cfor ‘ambient’ tests and−7�Cfor cold start tests and the soakperiods areusually12 h long. This requirement needs the efficient use of a climatically controlled buildingspace with temperature recording and where vehicles are not subjected to direct solarlight. There are several strategies that may be adopted by vehicle test facilities havingto deal with multiple vehicles in order to minimize the footprint of the soak area.

For cold start testing, vehicles, additional to the one rigged for the first test, maybe parked inside the chassis cell. They may then be moved on to the dynamometerby using special wheeled vehicle skids. ISO refrigerated shipping containers that canbe parked close to the cell access doors may be used to soak vehicles providingthe transit time, from container to cell, is as short as possible and without enginerotation. In ambient soak areas, it may be possible to use vertical stacking frames tostore cars two or three high and test them in order of access.

The European exhaust emissions test procedure

The new NECE urban driving cycle (UDC) was devised to represent city drivingconditions while the vehicle is on a chassis dynamometer. It is characterized by lowvehicle speed, low engine load, and low exhaust gas temperature and is used in thistext as an example of typical legislative drive cycles.

The UDC is used for emission certification of light duty vehicles in Europe.Before the test, the vehicle is allowed to soak for at least 6 h at a test temperature of20–30�C. The entire cycle includes four ECE segments, Fig. 16.9, repeated withoutinterruption, followed by one EUDC segment, Fig. 16.10.

Time, s

120

100

80

60

40

20

00 50 100 150 200

Speed,

km

/h

Figure 16.9 ECE 15 cycle

348 Engine Testing

120

100

80

60

40

20

0

Time, s

0 50 100 150 200

Sp

eed,

km

/h

Figure 16.10 EUDC cycle

The EUDC (extra urban driving cycle) segment has been added after the fourthECE cycle to account for more aggressive, high speed driving modes. The maximumspeed of the EUDC cycle is 120 km/h. An alternative EUDC cycle for low-poweredvehicles has been also defined with a maximum speed limited to 90 km/h.

There have been criticisms of these test procedures because of their lack ofseverity, in particular of the modest acceleration rates and it has been claimed thatthey underestimate emissions by 15–25 per cent compared with more realistic drivingat the same speed.

The US federal light duty exhaust emission test procedure(FTP-75)

This is a more complex procedure than the European test and is claimed to morerealistically represent actual road conditions. The cycle is illustrated in Fig. 16.11and, in contrast to the European test, embodies a very large number of speed changes.The cycle has three separate phases:

1. a cold-start (505-s) phase known as bag 1;2. a hot-transient (870-s) phase known as bag 2;3. a hot-start (505-s) phase known as bag 3.

The three test phases are referred to as bag 1, bag 2 and bag 3 because exhaustsamples are collected in separate sample bags during each phase. During a 10-mincooldown between the second and third phase, the engine is switched off. The 505-sdriving sequences of the first and third phase are identical. The total test time for theFTP 75 is 2457 seconds (40.95min), the top speed is 56.7mph, the average speed is21.4mph and the total distance covered is 11 miles.


00

70

60

50

40

30

20

10

240

Vehic

le s

peed, m

ph

480 720 960 1200

Test time, s

10 min soak

Hot startCold start

Bag 3Bag 2Bag 1

1440 1680 1920 2160 2400

Figure 16.11 EPA urban dynamometer driving schedule

Other test procedures

The US procedure (1985) for testing heavy-duty engines prescribes a transient testcycle, while the European (ECE 49) and Japanese (13-mode) procedures use steadystate tests. All three procedures measure emissions with the engine removed from thevehicle and installed on a dynamometer. Results are reported in g/bhp/h or g/kWh.

The emissions test procedures for motor cycles and mopeds in force in the USAand in Europe are essentially similar to those described above for passenger cars.Testing is carried out on a single roll dynamometer with inertia simulation, with aclamp to hold the motorcycle upright.

Vehicle evaporative emissions

Most vehicle emissions are produced by the combustion of hydrocarbon-based fuelswithin internal combustion engines; the exception that is important to the test engineeris the evaporative emissions of a complete vehicle.

These emissions arise because the vehicle both at rest and in motion containsvolatile compounds that evaporate differentially under changing solar and atmo-spheric conditions. In the 1970s, vehicles parked in the sun produced significantamounts of evaporative pollution. This pollution was quantified using a gastightchamber in which the vehicle was placed and subjected to typical temperature varia-tions, while exchanging the polluted air within the chamber with clean air from out-side (volume compensation). The resulting test methodology and legislation requiressuch a chamber, commonly referred to as a SHED (sealed housing for evaporativedetermination). These are usually purchased as complete integrated systems that aredesigned to meet legislation.

Typically, the SHED consists of a ‘box’ fitted with a vehicle access door capable ofbeing closed with a gasproof seal and lined with stainless steel and constructed from

350 Engine Testing

materials that are chemically inert at the temperatures used in testing. This box or con-tainer together with its external control cabinet is invariably installed within a largerbuilding and connected to the electrical and ventilation services of that building. Atypical car SHED unit will measure 6.5m long by 3m wide and 2.5m high and will beequipped with systems to ensure the safe operation while handling potentially explo-sive gases. Legislation, such as that of the California Air Resources Board (CARB),requires the test vehicle to be subjected to a diurnal temperature range of 20 and 35�C(98/69/EC) or 18.3 and 40.6�C, but some available units extend that range. Special,smaller than standard, SHED units for testing of motor cycles are available. There arealso special units that allow refuelling to be carried out within the sealed environment.

The withdrawal of polluted air for analysis and its replacement with treatedair in the controlled environment of the SHED requires special control and someexplosion-proof air handling equipment.

Major legislation currently requiring SHED testing includes

USA (EPA and CARB)

Hot soak test (40 CFR 86.138-96)Hot soak evaporative emission test (CARB)Diurnal evaporative emission test (40 CFR 86.133-96)Diurnal evaporative emission test (CARB)

Europe (European Union EU2000)

Hot soak test (Directive 98/69/EC)Evaporative emission test (Directive 98/69 EC)

Implications of choice of site on low level emissions testing

Test engineers not previously concerned with emissions testing should be awareof a number of special requirements that may arise, or have the same importance,elsewhere in other engine or vehicle testing. The first of these is the subject of thetest cell site. The problem with measuring exhaust emissions to the levels requiredto meet the standards set by SULEV is that the level of pollution of the ambient airmay be greater than that required to be measured at the vehicle exhaust. Almost anytrace of pollution of the air ingested by the engine or used to dilute the concentrationof exhaust gas within the measuring process, from exhaust gas of other engines orfumes from industrial plant, e.g. paint and solvents, will compromise the results ofemissions tests. Cross-contamination via ventilation ducts must be avoided in thedesign of the facility and it will be necessary to take into account the direction ofprevailing winds in siting the cell. Other than siting the facility in an area having, asfar as is possible, pristine air, there are three strategies available to help deal withSULEV level measurement:

1. Reduction of the exhaust dilution ratio by having a heated CVS system andemissions system which prevent water condensation.


2. Refinement or cleaning of the dilution air to reduce the background level ofpollutants to <0.1 p.p.m. This supports a CVS system, but is expensive and theplant tends to be large and complex.

3. Use a partial flow system with ‘pure’ (artificial) air for dilution. These bagmini-dilution systems do not comply with current CVS requirements and requireincorporation of an exhaust flow-rate sensor.

Time will tell as to which of these methods become preferred and legislative practice,meanwhile investment in emission laboratories sited in areas with high levels of airpollution would not seem to be advisable.

In addition to the siting of emission laboratories, the following items and trendsshould be considered by engineers involved in the design and development of testfacilities carrying out exhaust emission analysis:

• Since CO2 reduction requires fuel consumption reduction, the measurement of fuelconsumption will become more critical both in terms of mass fuel consumptionaccuracy and transient flow. Primary and secondary systems will have to beoptimized to obtain the best control and measurement accuracy.

• Liquid fuel conditioning (see Chapter 7) will become considerably more demand-ing. It will be necessary to take great care in the production, storage and handlingof fuel and in the avoidance of cross-contamination and deterioration. Control offuel temperature to better than ±1�C at the entry to the engine fuel rail is alreadyimportant and will continue to be so.

• It will become more necessary to condition combustion air, calling for the kindof systems described in Chapter 5. Temperature control of air will be required asa minimum with humidity control becoming more common and requiring a highstandard of system integration.

• Since much emissions testing will embrace cold starting and running at wintertemperatures there will be an increased need for cells capable of operating attemperatures of −10�C or lower.

• The requirement to test engines fitted with the full vehicle exhaust system willincrease the average cell floor area and call for exhaust dilution and extractionducts (see Chapter 6).

• The running of increasingly complex emission test cycles calls, in many cases,for expensive four quadrant dynamometers and elaborate control systems.

• The requirement for testing onboard diagnostic (OBD) vehicle systems willincrease.

The delicate and complex nature of much emissions measurement instrumentationwill call for standards of maintenance, cleanliness and calibration at a level moreoften associated with medicine than with automotive engineering.

352 Engine Testing

Notation

Air to fuel ratio AFRClean Air for Europe (EEC program) CAFÉCompression ignition (engine) CICompressed natural gas CNGCritical flow orifice CFOContinuously regenerating trap (diesel particle filter) CTRConstant volume sampling CVSDirect injection DIReference fuel with a total aromatic content of × wt-% DIR-xDiesel particulate filter DPFDiesel particulate matter DPMElemental carbon ECEnvironmental Protection Agency (USA) EPAHeavy duty HDIndirect injection IDILight duty LDOrganic carbon (bound in hydrocarbon molecules) OCHoriba analyser range trade name MEXA™Polyaromatic compounds PACPolyaromatic hydrocarbons PAHParticles in the size below 10�m PM10Soluble organic fraction SOFToxic equivalence factor TEF

Reference

1. Air Pollution from Motor Vehicles: Standards and Technologies for Controlling Emissions,World Bank (ISBN: 0821334441).

Further reading

BS 1747 Parts 1 to 13 Methods for Measurement of Air Pollution.BS 4314 Part I Infra-red Gas Analysers for Industrial Use.BS 5849 Method of Expression of Performance of Air Quality Infra-red Analysers.EEC L242 The Motor Vehicles (Type Approval) (Amendment) (No. 2) Regulations

1991 (http://www.legislation.gov.uk/si/si1991/Uksi_19912681_en_1.htm)EPA CFR 40–86 (http://www.access.gpo.gov/nara/cfr/waisidx_04/40cfr86_04.html)Horiba (1990) Fundamentals of Exhaust Emissions Analysis and their Application,

Horiba Instruments Ltd, Northampton.


Useful addresses

National Environment Research Council, Polaris House, North Star Avenue, Swin-don, Wiltshire, SN2 1EU, UK

Environmental Protection Agency, 401M Street SW, Washington DC 20460, USA.

17 Tribology, fuel and lubricationtesting

Introduction

Fuels and lubes (F&L) testing is a specialized branch of the engine test industry oftenrepresented by publications full of what are to the outsider impenetrable acronymstogether with changing roles for national, international and manufacture-specificorganizations.

F&L work is supported by laboratories analysing the chemical and physical char-acteristics of their products and parts of engines before and after strictly prescribedtests carried out either on test rigs or engine test cells. The oil industry has had toreact to, and support, engine developments aimed at reducing gaseous emissions; lowsulphur diesel and unleaded petrol being obvious examples of product development.There is a continuous development of fuels and lubricants required to prevent thepoisoning of more complex catalytic exhaust gas after treatment, while providinglonger periods between vehicle oil changes.

Lubricant classification and certification

Up to the early 1980s, the US military played a major role in the development of fueland lubricant specifications, but since that time and with the proliferation worldwideof interested parties, engine manufacturers, oil companies and chemical industry,a degree of harmonization of test methods has been achieved. Any student of thesubject will have to be aware of the acronyms or names of some of the governing orstandard setting organizations:

ACEA, the European Automobile Manufacturers AssociationASTM, the American Society for Testing of MaterialsATIEL, Association Technique de l’Industrie Européenne des LubrifiantsBSI, British Standards InstitutionCEC, Coordinating European Council – for the Development of Performance Tests

for Transportation Fuels, Lubricants and Other FluidsCOFRAC, the French Committee for Accreditation

Tribology, fuel and lubrication testing 355

DIN, Deutsches Institut für Normung (the German Institute for Standardization) isthe German national organization for standardization and is that country’s ISOmember body

EELQMS, European Engine Lubricants Quality Management SystemIP, Institute of PetroleumISO, International Organization for Standardization is the international standard-

setting body composed of representatives from national standards bodies

In Europe, the ACEA details lubrication oil specifications backed by EELQMSstandardized tests and performance criteria that have to be met by any product soldto their members. In addition to the ACEA tests, individual member companieshave their own test sequences and processes. The nomenclature and ACEA processis rather complex. In descending hierarchy oils are named by sequence, class andcategory plus a year of implementation designation. Some of the acronyms requiredto understand the documentation concerning classifications are

DPF Diesel particulate filterHTHS High temperature and high shear rate viscositySAPS Sulphated ash, phosphorus, sulphur (as in ‘low SAPS’)SCR Selective catalytic reduction (NOx control)TWC Three-way catalyst

The current ACEA sequences and classes are: ACEA A/B (combined) for gasolineand light duty diesel engines and ACEA C for gasoline and diesel engines havingadvanced after-treatment (catalyst compatible).

The category indicates the type of application the lubricant is intended for withinits class, there are currently four in the general sequence designated A1/B1, A3/B3,A3/B4 and A5/B5 and three in the catalyst compatible oils, C1 (low SAPS oils), C2(for vehicles with TWC and DPF) and C3 (high performance gasoline or diesels).

There are four classes in the oils designed for heavy duty diesels designated E2(general purpose, normal drain intervals), E4 (high rated engines meeting Euro 3 and4), E6 highly rated engines running on low SAPS fuel) and E7 (highly rated engineswith EGR and SCR).

The year of implementation number is placed after the class designation thus:C1-04

Using these classes to set the limits of laboratory- and engine-based tests, theACEA produces tables indicating the minimum allowable quality of oils of thevarious classes under a system of self-assessment to EELQMS.

Tribology

The science of tribology, launched as a separate discipline with a new name by theJost Report of 1966, is concerned with lubrication and wear; both are central to i.c.engine operation. It was pointed out some years ago that the difference between a

356 Engine Testing

brand new road vehicle and one that was totally worn out was the loss of perhaps100 g of material at critical points in the mechanism. As far as friction is concerned,the great majority of the power developed by an automobile engine is ultimatelydissipated as friction. Even a large slow-speed diesel engine is unlikely to achievea mechanical efficiency exceeding 85 per cent, most of the losses being due tofriction.

As evolution proceeds, different aspects of engine lubrication and wear call forconcentrated attention. Some problems, such as excessive bore wear and bore pol-ishing in diesel engines, are largely overcome and sink into the background whileothers emerge as engine technologies such as common-rail diesel develop, operatingconditions become more demanding or emission legislation sets new parameters.The following may be regarded as the main tribology-related problems of particularcurrent concern:

• valve train wear;• injector life;• control of lubricant consumption;• interactions between fuel and lubricant;• development of synthetic and ‘stay in grade’ or ‘fill for life’ lubricants;• piston design, with particular reference to top land clearance, ring wear, oil

control, distortion and reduction of friction;• the effect on exhaust emissions of lubrication oil constituents.

The pursuit of higher efficiency and reduced friction losses is a continuing activityin all engine development departments.1�2

Bench tests

A wide range of standard (non-engine) tests are written into standards such as thoseof the ASTM, DIN, CEC and IP.

These tests for oils can range from chemical analysis such as ASTM D 1319(aromatics and olefins content), through physical laboratory tests such as ASTM D2386 (freezing point), to special rig tests such as DIN 51 819-3 which uses a rollerbearing test rig. Similarly, there are series of tests specifically for greases and fuelanalysis.

The principle of operation of many of the lubrication oil test machines has notchanged for many years but has been updated because, in spite of the fact thatthe test conditions they impose differ from any to be found in a real engine, theirvalue is their ability to reliably reproduce the classic mechanical wear and/or failureconditions.

The number of analytical and diagnostic tools available to test laboratories(Table 17.1) has increased significantly in the last two decades and range from laser-measuring devices able to differentiate into the realms of nanotechnology, throughuse of radiotracing techniques of wear particles, to electronic scanning microscopes.


Table 17.1 Tribology: surface examination techniques

Visual and microscopic examinationWear debris quantifying and analysisSurface roughness measurementMetallographical sections and etchingMicro-hardness measurementScanning electron and scanning tunnelling electron microscopyInfrared and Raman spectroscopy

After each engine test, whether it is oil- or fuel-related test sequence, the wholeengine or that part of it of specific interest will be stripped under laboratory conditionsand examined. In addition to the surface examinations described above which lookat discolouration, corrosion and wear, fuel tests will examine such effects as thereduction in flow through injector nozzles and ring sticking.

Oil characteristics

Viscosity is a measure of the resistance of a fluid to deform under shear stress.It is commonly perceived as resistance to pouring. It is measured by a viscometerof which there are various designs commonly based on the time taken to flowthrough an orifice or a capillary tube when at a standard temperature. The SAE‘multigrade’ rating known to most motorists has a nomenclature based on flowthrough an orifice at 0�F and 210�F, with ‘W’ (standing for winter) after the numberfor low temperature rating and a plain number for the high temperature rating as inSAE10W40.

Dynamic viscosity

The SI physical unit of dynamic viscosity � is the pascal-second (Pa·s), which isidentical to 1 kg·m−1·s−1

The physical unit for dynamic viscosity in the cgs∗ is the poise (P) namedafter Jean Poiseuille. It is more commonly expressed, particularly in ASTM stan-dards, as centipoise (cP). The centipoise of water is almost unity (1.0020 cP)at 20�C.

1 poise = 100 centipoise = 1 g·cm−1·s−1 = 0.1 Pa·s.1 centipoise = 1 mPa·s.

∗ Centimetre gram second system of units.

358 Engine Testing

Kinematic viscosity

Kinematic viscosity is the ratio of the viscous force to the inertial force or fluiddensity �.

� =�

�

Kinematic viscosity has SI units (m2·s−1). The cgs physical unit for kinematic vis-cosity is the stokes (St), named after George Stokes. It is sometimes expressed interms of centistokes (cS or cSt).

1 stokes = 100 centistokes = 1 cm2·s−1 = 0.0001 m2·s−1

Viscosity index (or VI) is a lubricating oil quality indicator, an arbitrary measurefor the change of kinematic viscosity with temperature.

Total base number (TBN)

The TBN of oil is the measure of the alkaline reserve, or the ability of the oil toneutralize acids from combustion tested to ASTM D-2896 by a laboratory methodknown as potentiometric perchloric acid titration. Depletion of the TBN in serviceresults in acid corrosion and fouling within the engine.

Appendix 17.1, 17.2 and 17.3 show examples of the standards and test methodsof the properties of lubricating oils, diesel fuels and gasoline, respectively; for fullup to date listings, the websites maintained by the relevant authority is recommendedas a starting point.

Reference fuels and lubricants

Any laboratory carrying out fuels and lube testing will have to store and handlecorrectly reference fuels and oils. Many component inspection results after runningengine tests will be based on comparative judgements based on the use of thesereference liquids which are expensive and may have limited storage life unless keptin optimum conditions. Connection of reference fuels into test cells is covered inChapter 7.

Test limits for such criteria as piston cleanliness or sludge accumulation in someengine-based lubrication tests will be based on reference lubricant results that arecommonly designated with a CEC reference, such as in the case of CEC RL191(15W-40).


Designated engines and test regimes in fuel and lube testing

Whereas tribology studies the phenomena associated with wear and lubrication, theclassification and specification of all fuels and crankcase lubricants worldwide arebased upon standardized engine tests.

There is a complete section of the engine test industry attached to the oil industryusing internationally specified engines in exactly determined configurations to runstrict test protocols in order to certify engine fuels and lubricants. This use of aprescribed engine builds redundancy and the need for regular updating into thecertification process, since the life and availability of any engine is limited as is itssuitability to represent the current technology that the fuels and lube industry has tosupport. At the time of writing, the PSA 2.1 turbo diesel engine XUD11 BTE that isused in several tests such as the 95 h duration lubricant test, CEC L-56-T-98, can bedifficult to obtain in some configurations and was due to be replaced in 2006 withthe DV4TD engine. Examples of just some of the engines commonly designated intests are:

• Mercedes M-111 used in the fuel economy test CEC-L-54-T-96 (C1-04, C2-04,C3-04);

• Mack T-8E used in the ‘soot in oil’ test ASTM 5967 (E4-99, E6-04, E7-04);• Peugeot TU5JP-L4 used in the 72-h piston and oil viscosity test CEC-L-88-T-02

(A1/B1-04, A3/B3-04, A3/B4-04, A5/B5-04).

The approved engines form part of the standard rig which is arranged specificallyfor dealing with that particular model.

A fuel and lube test stand will typically be configured specifically for a desig-nated engine that will remain in place for prolonged periods, only being removed inwhole or part for component inspection after testing. The engine will be mountedon a base-plate, connected to an eddy-current dynamometer and rigged with fueland/or oil control systems instrumented exactly as required by regulation. The run-ning of the test sequence, the manipulation of the test results and the post-testexamination of engine parts, such as valves and injectors, are similarly tightly pre-scribed.

The major disadvantages of such tests are the cost and time involved and the diffi-culty, where measurements of such factors as surface wear, deposits, oil consumptionand friction are concerned, of obtaining consistent results. The measurements canbe much affected by variations in engine build and components, also of fuel in thecase of lubrications or of lubrication oil in the case of fuel tests; therefore mostfired tests are run with strictly prescribed reference fluids. There are a few inter-nationally recognized specialist test laboratories that provide the service of runningsuch tests and each of the major engine manufacturers have their own special testcells running both public domain test sequences and their own tests on new enginedesigns.

Motored engine or ‘part engine’ tests are used for assessing specific componentssuch as valve trains. They are somewhat removed from real conditions, but can

360 Engine Testing

22.5%

19%

25%

12.5%

10%

6%5%

Figure 17.1 Distribution of friction losses in an i.c. engine

give a good insight into local problems of wear and lubrication. By the successiveremoval of components from a motored engine, a good estimate can be made of thecontribution made to frictional losses by the different components, see Fig. 17.1.

This book has mentioned some of the very numerous standard tests that are inconstant use for assessment and routine confirmation of fuel and lubricant properties.The development, updating and monitoring of these standard tests, not to mentionthe cost in fuel, components and man-hours, represent a major burden to engine andlubricant manufacturers.

Example of engine tests involving tribology

Measurement of compression ring oil film thickness

A feature of many tribological situations is the extreme thinness of the oil filmspresent, usually to be measured in microns, and various techniques have been devel-oped for measuring these very small clearances.3 Figure 17.2 shows measured oilfilm thicknesses between top ring and cylinder liner near top dead centre in a large


6

4

2Induction stroke

Compression stroke

0

Film

thic

kness (

µm

)

6

4

2

00 25 50 75 100 0 25 50 75 100

Exhaust strokeFiring stroke

Torque (%)SAE 15W/40

SAE 10W/30

Figure 17.2 Oil thickness between ring and cylinder liner during four-stokeoperation

turbocharged diesel engine. There is a wide variation in film thickness for the dif-ferent strokes and an effective breakdown during the firing stroke. It is notable thatthe thinner oil, SAE 10W/30, gives better protection than a thicker one. This wasthought to be the result of higher oil flow to this critical area with the thinner oil andis confirmed by the generally lower rates of wear of aluminium from the piston andchromium from the ring surfaces.

Wear measurements by irradiation

Figure 17.3 shows wear measurement using an irradiated piston ring in an experi-mental single cylinder diesel engine and indicates a steady wear rate following initialrunning-in. Wear rate was also found to be sensitive to oil viscosity in this test.This technique is perhaps the only one available for indicating very small incrementsof wear.

362

Engin

eTestin

g

30

20

100

(a)3

0

20

10

Al wear (p.p.m.) Cr wear (p.p.m.)

1000 rev/min, 743 N m

1300 rev/min, 806 N m

1600 rev/min, 743 N m

1800 rev/min, 728 N m

2000 rev/min, 679 N m

2200 rev/min, 765 N m

0

(b)

SA

E 1

0W

/30

SA

E 1

5W

/40

Figure

17.3

Rates

ofwear,

chromium

from

ringan

dalu

minium

from

pisto

n,

engin

eas

forFig.

17.2


Biofuels

In order to reduce dependency on fossil fuel oil stocks, there are a number of globalinitiatives intended to increase the use of oil derived from renewable sources (seeChapter 7 concerning storage of biofuels). The European Commission introducedDirective 2003/30/CE which proposes increasing the proportion of energy deliveredby biofuels to 5.75 per cent by 2010.

Biodiesel in Europe has to conform to the EN14214 specification and can bederived from a range of animal and vegetable oils, the most common forms are fattyacid methyl ester (FAME) and fatty acid ethyl ester (FAEE). The most commonsource of biodiesel in Europe is vegetable rape, from which rape-seed methyl ester(RME) is produced. RME is currently more expensive to produce, but has someperformance advantages over FAME the most critical of which is low temperatureperformance. RME biodiesel has a flash point of >150�C which is significantlyhigher than that of petroleum diesel.

Engine test laboratories will be engaged in testing the emissions and performanceof the many possible mixtures and concentrations of these bio-fossil fuels in thecoming years. The tendency for some bio-fuels to form wax crystals at low temper-atures will require new additives and filter flow testing∗∗ in climatic cells; a wholenew direction of testing is opening up.

It should be noted that there are worldwide environmental movements encouragingthe home production of biodiesel feedstock, the contents of which will be uncontrolledand the performance of which will be highly variable.

Biodiesel is reported to reduce emissions of carbon monoxide (CO) and carbon di-oxide, but it should be remembered that the carbon in biodiesel emissions is within theanimal ‘carbon cycle’ rather than being new carbon released from that tied up inmineraldeposits. They are also reported to contain fewer aromatic hydrocarbons but to pro-duce higher emissions of nitrogen oxide NOx than diesel from petroleum feedstock.

Ethanol (C2H6O) can be blended with gasoline in varying quantities; the resultingfuel is known in the United States as gasohol, or gasoline type C in Brazil wherethere is considerable experience in general use of E20 and E25.

E85 is being used in public service vehicles in the USA; the 15 per cent gasolinebeing required to overcome cold starting problems experienced with higher ethanolmixtures.

Methanol (CH3OH) can also be used as a gasoline additive, but is highly toxicand less easy to produce from biological sources than ethanol.

Summary

The fuels and lube test industry, which is a specialized branch of the test industry,has been briefly described. The point to be borne in mind by the test engineer who

∗∗ Cold filter plugged point (CFPP) is a standard test requirement for winter grade fuels.

364 Engine Testing

may be confronted with a problem of wear or friction is that it is unlikely to be solvedon the test bed alone. There exists a vast range of specialized tests and techniquesthat eliminate a great deal of expensive engine testing and in many cases can clarifyaspects of the problem that are simply not susceptible to investigation in the complexenvironment of a running engine.

For guidance in the tribological aspects of engine performance and testing, theTribology Handbook will be found valuable.4�5

The subject of biofuels will become very important to the fuel test and cer-tification industry as political pressure to reduce the dependency on fossil fuelsbuilds.

References

1. Crankcase Oil Specification and Engine Test Manual (1987) Paramins, Exxon Chemical,Technology Centre, Abingdon.

2. Friction and Wear Devices (1976) American Society of Lubrication Engineers.3. Moore, S.L. (1987) The effect of viscosity grade on piston ring wear and oil film thickness

in two particular diesel engines, Proc. I. Mech. E. C184/87.4. Neale, M.J. (ed.) (1993) Lubrication: A Tribology Handbook, Butterworth-Heinemann,

Oxford.5. Neale. M.J. (ed.) (1993) Bearings: A Tribology Handbook, Butterworth-Heinemann,

Oxford.

Further reading

Automotive Lubricants Reference Book, 2nd edn ISBN: 1-86058-471-3.BS 2000 Parts 0 to 364, Methods of Test for Petroleum and its Products.

Plint, M.A. and Alliston-Greiner, A.F. (1990) Routine engine tests: can we reducetheir number, Petroleum Review, July.

Appendix 17.1 Listing of major CEC tests and publications

L-02-A-78 Oil oxidation and bearing corrosionL-33-T-82 Biodegradation of two-stroke outboard engine

lubricantsL-38-T-87 Valve train scuffingL-41-T-93 Evaluation of sludge inhibition qualities of motor oils

in a gasoline engineL-07-A-85 Load carrying capacity of transmission lubricantsL-11-T-72 Coefficient of friction of automatic transmission

fluids


L-45-T-93 Viscosity shear stability of transmission lubricantsL-17-A-78 Cam and cylinder wear in diesel enginesL-24-A-78 Engine cleanliness under severe diesel conditionsL-42-A-92 Evaluation of piston deposits and cylinder bore polishingL-14-A-88 Mechanical shear stability of lubricants containing

polymersL-36-A-90 The measurement of lubricant dynamic viscosity under

conditions of high shear (Ravenfield viscometer)L-37-T-85 Shear stability test for polymer-containing oilsL-39-T-87 Oil/elastomer compatibility testL-40-T-87 Evaporative loss of lubricating oilsF-03-T-87 Evaluation of gasoline with respect to maintenance of

carburettor cleanlinessF-04-A-87 The evaluation of gasoline engine intake system depositsF-05-T-92 Intake valve cleanlinessM-02-A-78 Internal combustion engine rating procedureM-07-T-83 The relationship between knock and engine damageM-08-T-83 Cold weather drivability test procedureM-09-T-84 Hot weather drivability test procedure

M-10-T-87 Intake system icing test proceduresM-11-T-91 Cold weather performance test procedure for diesel

vehiclesM-12-T-91 Representative sampling in service of marine crankcase

lubricantsM-13-T-92 Analysis of marine crankcase lubricantsP-108-79 Manual of CEC reference/standardization oils for

engine/rig testsP-221-93 A guide to the definition of terms relating to the viscosity

of engine oilsP-222-92 CEC reference fuels manual, 1992, annual reportL-18-U-93 Low temperature apparent viscosityL-12-U-93 Piston cleanliness

Authority responsible for tests in Europe

The Coordinating European Council for the Development of Performance Tests forLubricants and Engine Fuels (CEC)Madou Plaza, 25th FloorPlace Madou 1B 1030 BrusselsBelgium

366 Engine Testing

Appendix 17.2 Properties of gasoline: standards andtest methods

BSI

BS 2000-12 Methods of test for petroleum and its products

ASTM

D2623-83 Knock Characteristics of Liquefied Petroleum (LP) Gases bythe Motor (LP) Method

D2699-84 Knock Characteristics of Motor Fuels by the ResearchMethod

D2700-84 Knock Characteristics of Motor and Aviation-Type Fuels bythe Motor Method

D2886-83 Knock Characteristics of Motor Fuels by the DistributionOctane Number (DON) Method

D2885-84 Research and Motor Method Octane Ratings Using On-LineAnalyzers

D439-85a Automotive GasoleneD3710-83 Boiling Range Distribution of Gasolene and Gasolene

Fractions by Gas ChromatographyD240-85 Heat of Combustion of Liquid Hydrocarbon Fuels (General

Bomb Method)D2551-80 Vapor Pressure of Petroleum Products (Micromethod)D323-82 Vapor Pressure of Petroleum Products (Reid Method)D2889-81 Vapor Pressures (True) of Petroleum Distillate

FuelsD93-85 Flash Point by Pensky-Martens Closed

Tester

Institute of Petroleum

IP34/80 Flash Point by Pensky Martens ClosedTester

IP325/80 Front End Octane Number (RON 100�C) of MotorGasolene

IP12/73 Heat of Combustion of Liquid HydrocarbonFuels

IP236/69 Knock Characteristics of Motor and Aviation Type Fuels bythe Motor Method

IP237/69 Knock Characteristics of Motor Fuels by the ResearchMethod

IP15/67 Pour Point of Petroleum OilsIP171/65 Vapour Pressure – Micro MethodIP69/78 Vapour Pressure – Reid Method


Appendix 17.3 Properties of diesel fuels: standards andtest methods

BSI

BS 2000-12 Methods of test for petroleum and its productsBS 2869 Fuel oils for oil engines and burners for non-marine

useBS MA 100 Petroleum fuels for marine oil engines and boilers

ASTM

D613-84 Ignition quality of diesel fuels by the cetane methodD975-81 Diesel fuel oilsD976-80 Cetane index, calculated, of distillate fuelsD189-81 Carbon residue, Conradson, of petroleum productsD524-81 Carbon residue, Ramsbottom, of petroleum

productsD2500-81 Cloud point of petroleum oilsD93-85 Flash point by Pensky Martens closed testerD240-85 Heat of combustion of liquid hydrocarbon fuels

(general bomb method)D97-85 Pour point of petroleum oilsD3245-85 Pumpability test for industrial fuel oilsD129-64 Sulfur in petroleum products by the bomb methodD4294-83 Sulfur in petroleum products by nondispersive

X-ray fluorescence spectrometryD1552-83 Sulfur in petroleum products (high-temperature

method)D1266-80 Sulfur in petroleum products (lamp method)D2709-82 Water and sediment in distillate fuels by centrifuge

Institute of Petroleum

IP218/67 Calculated cetane index of diesel fuelsIP13/78 Conradson carbon residue of petroleum productsIP34/80 Flash point by Pensky Martens closed testerIP12/73 Heat of combustion of liquid hydrocarbon fuelsIP41/60 Ignition quality of diesel fuelsIP15/67 Pour point of petroleum oilsIP14/65 Ramsbottom carbon residue of petroleum productsIP107/73 Sulphur in petroleum productsIP276 Total base number

18 Chassis or rolling roaddynamometers

Introduction

This chapter deals with an area of testing activity that is closely allied to the testingof automotive engines and which increased enormously in importance during thelast quarter of the twentieth century. It is exclusively concerned with road vehiclesand their power-train systems and its expansion has been driven by a number ofdevelopments:

• Technical: the need to perfect the complex engine, transmission and whole vehiclemanagement systems that are demanded by the modern road user, the economicaland repeatable performance of endurance testing, investigations of noise, vibration

Chassis or rolling road dynamometers 369

and harshness (NVH), emissions testing, checking performance under extremeclimatic conditions, etc.

• Legislative: the need to comply with ever more exacting statutory requirementsgoverning both new vehicles and the after-market in such areas as vehicle safety,brake performance, NVH, electromagnetic compatibility (EMC), fuel consump-tion and dominating all other tasks, exhaust emissions.

These requirements have called into existence a hierarchy of dynamometers and testchambers of increasing complexity:

• brake testers;• in-service tuning, ECU remapping;• end-of-line production testing;• emissions certification/testing installations;• mileage accumulation facilities associated with emission testing laboratories;• anechoic and NVH and EMC test chambers;• chassis dynamometers within climatic chambers and wind tunnels;• independent wheel dynamometers.

The major types and uses are discussed later in this chapter.

The road load equation

The behaviour of a vehicle under road conditions is described by road load equations(RLE). The RLE is a fundamental requirement of a true chassis dynamometer whichhas to resist the torque being produced at the vehicle drive wheels in such a wayas to simulate ‘real-life’ resistance to vehicle motion. It is necessary to have a clearunderstanding of this control model before studying the various types of machinescoming under the heading of chassis dynamometers or rolling roads.

Increasing use of transient testing and road simulation in modern engine testcells means that the following section is relevant to engine test students andengineers.

The RLE is the formula that calculates the change in torque required with change ofvehicle speed in order to simulate the real-life performance of a given vehicle; itdefines the traction or braking force that is called for under all possible conditions,which include

• steady travel at constant speed on a level road;• hill climbing and descent;• acceleration, over-run and braking;• transitions between any of the above;• effects of atmospheric resistance, load, towed load, tyre pressure, etc.

370 Engine Testing

The road load equation for a given vehicle defines the tractive force or retardingforce, F (Newton), that must be applied in order to achieve a specified response tothese conditions. It is a function of the following parameters:

Vehicle specific:Mass of vehicle M kgComponents of rolling resistance a0 NSpeed dependent a1V NAerodynamic a2V

2 N

External:Speed V m/sRoad slope � rad

The usual form of the road load equation is:

F = a0+a1V +a2V2+MdV/dt+Mg sin � (1)

where M dv/dt = force to accelerate/brake vehicle and Mg sin � = hill climbingforce.

More elaborate versions of the equation may take into account such factors aswheel spin and cornering.

The practical importance of the road load equation lies in its application to thesimulation of vehicle performance. It forms the link between performance on theroad and performance in the test department.

To give a feel for the magnitudes involved, the following equation relates to atypical four-door saloon of moderate performance, laden weight 1600 kg:

F = 150+3V +0�43V 2+1600dV/dt+1600×9�81 sin �

The various components that make up the vehicle drag are plotted in Fig. 18.1awhich shows the level road performance and makes clear the preponderant influenceof wind resistance.

The road load equation predicts a power demand at the road surface of 14.5 kWat 60mph, rising to 68.8 kW at 50m/s (112mph) for a level road.

These demands are dwarfed by the demands made by hill climbing and accel-eration. Thus, Fig. 18.1b, to climb a 15 per cent slope at 60mph, calls for atotal power input of 14.5+ 63.2= 77.7 kW while, Fig. 18.1c, to accelerate from 0to 60mph in 10 s at constant acceleration, calls for a maximum power input of14.5+ 115= 129.5 kW.


1000

500

N

00

(a) Level road, drag

10 20

m/s

30 40 50

Constant

Speed dependent

Wind resistance

(b) Hill climbing force

5000

2500

0

N

Slope (%)

5 10 15 20

N

7500

5000

2500

014 12 10

Time, 0–60 mph

(c) Acceleration force

8 6s

Figure 18.1 Vehicle drag: saloon of laden weight 1600 kg (a) level road per-formance; (b) hill climbing; (c) acceleration force

Genesis of the rolling road dynamometer

The idea of running a complete vehicle under power while it was at rest was firstconceived by railway locomotive engineers before being adopted by the road vehicleindustry. As a matter of historical record the last steam locomotives built in the UK

372 Engine Testing

were tested on multiple roll units, with large eddy-current dynamometers∗ connectedto each driven roll, the tractive force being measured by a mechanical linkage andspring balance. To an observer close to the locomotive on the test rig it must havepresented an awe-inspiring sight, running at full power and a wheel speed of 90mph.

Today the chassis dynamometer is used almost exclusively for road vehicles,although there are special machines designed for fork-lift and articulated off-roadvehicles. The advantages to the designer and test engineer of having such facilitiesavailable are obvious; essentially, they allow the static observation and measurementof the performance of the complete vehicle while it is, in most respects, in motion.

Before the 1970s, most machines were comparatively primitive rolling roads,characterized by having rollers of rather small diameter, which inadequately simulatedthe tyre contact conditions and rolling resistance experienced by the vehicle on theroad and fitted with various fairly crude arrangements for applying and measuringthe torque resistance. A single fixed flywheel was commonly coupled to the rollerto give a simulation of a vehicle inertia.

The main impetus for development came with the rapid evolution of emissionstesting in the 1970s. The diameter of the rollers was increased, to give more realistictraction conditions, while trunnion mounted d.c. dynamometers with torque measure-ment by strain gauge load cells and more sophisticated control systems permittedmore accurate simulation of road load. The machines were provided with a rangeof flywheels to give steps in the inertia; precise simulation of the vehicle mass wasachieved by ‘trimming’ the effective inertia electrically.

In recent years the development in digital control techniques and electrical driveshas meant that incremental flywheels have been dispensed with in most emissiontesting designs and been replaced by single (2× 2) or double (4× 4) roll machinesreliant on electrical simulation of all aspects of ‘road behaviour’.

Brake testers

These simple machines, commonly installed in service garages and government teststations, are used mainly for the statutory ‘in-service’ testing of cars and commercialvehicles. They are installed in a shallow pit, and consist essentially of two pairsof rollers, typically of 170-mm diameter and having a grit-coated surface to give ahigh coefficient of adhesion with the vehicle tyre. The rollers are driven by gearedvariable speed motors and two types of test are performed. Either the vehicle brakesare fully applied and torque increased until the wheels slip, or the rollers are drivenat a low surface speed, typically 2 km/h, and the relation between brake pedal effortand braking force is measured.

The testing of a modern ABS brake system requires a separately controlled rollerfor each wheel and a much more complicated interaction with the vehicle braking

∗ The National Railway Museum in York holds records and displays the Heenan & Froudespecification plate.


system. This calls for computerized control of the tester. In many vehicle managementsystems, a plug is provided for connection of the brake system to a suitable braketester computer. Modern ‘end of line’ brake test machines are used by vehiclemanufacturers in a series of final vehicle check stations and therefore have features,such as communication with the test vehicle control systems, that are model specific.

Rolling roads for in-service tuning and assessment

For the typical vehicle maintenance garage, the installation of a rolling road and itsinstallation with a building represents a considerable investment, only justified if thegarage has a good market for its expertise in vehicle tuning or other specialized skills.

To cater for this market a number of manufacturers produce complete installationsbased on small diameter rollers, thus requiring minimum subfloor excavation. Thismarket has tended to be dominated by American suppliers and grew rapidly whenthe US Environmental Protection Agency (EPA) called for annual emissions testingof vehicles, based on a chassis dynamometer cycle (IM 240).

Road load simulation capability at this level is usually limited to that possiblewith a choice of one or two flywheels and a comparatively simple control system;most of the testing is concerned with optimizing (tuning) the engine managementsystem and straightforward maximum power certification. The tests have to be ofshort duration to avoid overheating of engine and tyres.

In-service rolling roads for testing large trucks are confined almost exclusively tothe USA. They are based on a single large roller capable of running both single- anddouble-axle tractor units. The power is usually absorbed by a portable water brake,such as that shown in Fig. 4.1, which may also be used for direct testing of engines,a useful facility in a large agency test facility.

Rolling roads for end-of-line production testing

Test rigs for this purpose range from simple roller sets to multi-axle units withinterconnected roll sets and road load simulation. Certain special design features arerequired and any specification for such an installation should take them into account:

• The design and construction of the machine must minimize the possibility ofdamage from vehicle parts falling into the mechanism. It is quite common forfixings to shake down into the rolling road, where they can cause damage if thereare narrow clearances or converging gaps between, for example, the rolls andlift-out beam, or in the wheel-base adjustment mechanism.

• The vehicle must be able to enter and leave the rig quickly yet be safely restrainedduring the test. The usual configuration is for all driven wheels to run between tworollers. Between each pair of rolls there is a lift-out beam that allows the vehicle toenter and leave the rig. When the beam descends it lowers the wheels between the

374 Engine Testing

rolls and at the same time small ‘anti-climb-out’ rollers swing up fore and aft ofeach wheel. At the end of the test, with the wheels at rest, the beam rises, therestraining rollers swing down, and the vehicle may be driven from the rig.

• To restrain the vehicle from slewing from side to side to a dangerous extent,specially shaped side rollers are positioned between the rolls at the extreme widthof the machine where they will make contact with the vehicle tyre to preventfurther movement. These rollers can cause problems if a wide variety of tyreprofiles are used on test vehicles particularly in the case of low profile tyres.

• Particular care should be taken with the operating procedures if ‘rumble strips’are to be used. These are raised strips or keys sometimes fitted to the rolls inline with the tyre tracks, their purpose being to excite vibration to assist in thelocation of vehicle rattles. However, if the frequency of contact with the tyreresonates with the natural frequency of the vehicle suspension this can give riseto bouncing of the vehicle and even to its ejection from the rig. They can alsogive rise to severe shock loads in the roll drive mechanism.

Rigs designed to absorb power from more than one axle may have to be capableof adjustment of the centre distance between axles. The complexity of a shaft orbelt connection between the front and rear roller sets has been obviated by modernelectronic speed synchronization. Special purpose roller rigs for checking wheelalignment sometimes form part of the end-of-line test equipment.

Chassis dynamometers for emissions testing

The majority of chassis dynamometers manufactured worldwide in the last 25 yearshave been used for some form of emissions testing or homologation.

The standard emissions tests developed in the USA in the 1970s were based on arolling road dynamometer developed by Clayton and having twin rolls of 8.625-inch(220mm) diameter. This machine became a de facto standard despite its limitations,the most serious of which was the small roll diameter, which resulted in tyre contactconditions much different from those on the road.

Later models used rollers of 500mm diameter coupled to a four-quadrant d.c.dynamometer with declutchable flywheels to simulate steps of vehicle inertia withinwhich gaps the electrical power operated, Fig. 18.2.

The machine is designed for roller surface speeds of up to 160 km/h and the digitalcontrol system is capable of precise road load control, including gradient simulation.Such machines must be manufactured to a very high standard to keep vibration to anacceptable level. It is possible to abuse them, notably by violent brake applicationwhich can induce drive wheel bounce and thus very high stresses in the roller andflywheel drive mechanism.

The US Environmental Protection Agency (EPA), which remain the authoritydefining chassis dynamometers of this class worldwide, have approved machineshaving a single pair of rollers of 48-inch diameter and 100 per cent electronic inertia


(A) (A)

(A)

(D)

(E)(A)

(C)

(C)

B

Figure 18.2 Schematic plan view of four roller (2×2= single driven axle)emission chassis dynamometer of late 1980s design. (A) Four 500-mm rollers;(B) selectable flywheel set; (C) belt drive; (D) d.c. motor fitted with shaftencoder; (E) disk brake for vehicle loading

simulation. A number of standard designs are now available from the major suppliersand most are of a compact design having a double-ended a.c. drive motor with rollersmounted overhung on each stub shaft. This type of design can produce a standard,pretested package that enables chamber installation time to be minimized.

Mileage accumulation facilities

As a direct result of emissions regulations and the practice of homologation, it hasbecome necessary for manufacturers to check on the change and probable deteriorationin emissions performance after the vehicle has been driven for up to 80 000 km. Onecurrent legislative trend is to increase the length of this certified ‘useful life’ of thevehicles emission control system from 50 000 to 100 000 miles or greater.

The cost and the physical strain of using human drivers on test tracks or publicroads for driving vehicles the prescribed distances in as little time as possible arehigh so special chassis dynamometer systems have been developed for running thespecified sequences under automatic control, commonly for a period extending to 12weeks.

Mileage accumulation dynamometers comprise a single large roll, of 48 inches orlarger diameter, directly coupled to a d.c. or a.c. motor having sufficient power torun the required repetitive test sequence; for the average saloon car a motor capacityof 150 kW is sufficient.

In order to fully automate the process, the test vehicle is usually fitted with a robotdriver. A bewildering range of these devices is available on the world market: for

376 Engine Testing

this particular application reliability must be the prime consideration. Some devicescan be mounted on the driver’s seat and some are capable of also operating thefootbrake and engaging reverse gear, but these refinements are not required by currentlegislative drive cycles. The set-up time of the robot may be appreciable but shouldbe assessed against a test duration of 12 weeks or more. If wear or deterioration takesplace in the robot mechanism or in the vehicle controls, for example by change in theclutch ‘bite’ position, the control system should be capable of automatic rectification,usually by means of periodic ‘relearn’ and adjustment cycles.

Due to the obvious problems of ventilating such a facility it is invariably locatedoutside, under a simple roof. The control room housing the computerized control andsafety systems can occupy a small building at one end of the facility, which in somecases may comprise up to 10 chassis dynamometers.

As the vehicle’s own cooling system is stationary, a motorized cooling fan, facingthe front grill, is essential. The fan will be fitted with a duct to give a reasonablesimulation of air flow under road conditions, and must be firmly anchored. The fanspeed is usually controlled to match apparent vehicle speed up to about 130 km/h;above that speed noise and fan power requirements become an increasing problem.Since mileage accumulation facilities usually run 24 h a day, they must be suitablyshielded to prevent noise nuisance.

Automatic refuelling systems may be fitted in the facility for each chassisdynamometer unit.

Special purpose chassis dynamometers

Noise, vibration and harshness (NVH), electromagnetic compatibility (EMC)rigs and anechoic chamber dynamometers

Some features of anechoic cell design are discussed in Chapter 3. A critical require-ment is that the chassis dynamometer should itself create the minimum possible noise.The usual specification calls for the noise level to be measured by a microphonelocated 1m above and 1m from the centreline of the rolls. Typically, the requiredsound level when the rolls are rotating at a surface speed of 100 km/h will be in theregion of 50 dBA.

To reduce the contribution from the dynamometer motor and its drive system, itis usually located outside the main chamber in its own sound-proofed compartmentand connected to the rolls by way of a long shaft (the design of these shafts canpresent problems; lightweight tubular carbon fibre shafts are sometimes used). Thedynamometer motor will inevitably require forced ventilation and the ducting willrequire suitable location and treatment to avoid any contribution of noise.

Hydrodynamic bearings on quiet roll systems are common, but not without theirown problems as their theoretical advantages in terms of reduced shaft noise maynot be realized in practice, unless noise from the pressurized supply oil system issufficiently attenuated.


For a well-designed chassis dynamometer, the major source of noise will be thewindage generated by the moving roll surface and this is not easily dealt with.Smooth surfaces and careful shrouding can reduce the noise generated by the rollerend faces, but there is inevitably an inherently noisy jet of air generated where theroller surface emerges into the test chamber. If the roller has a roughened surface tosimulate road conditions the problem is exacerbated. The noise spectrum generatedby the emerging roller is influenced by the width of the gap and in some casesadjustment is provided at this point.

The flooring over the dynamometer pit, usually of steel plate, must be carefullydesigned and appropriately damped to avoid resonant vibrations. A particular featureof NVH test cells is that the operators may require access to the underside of thevehicle during operation. This is usually by way of a trench at least 1.8m deep, lyingbetween the vehicle wheels.

Testing for electromagnetic compatibility (EMC)

Worldwide safety standards include regulations governing the acceptable levels ofelectromagnetic interference produced by road vehicles and the vulnerability ofvehicle electronic systems to powerful directional beams of electromagnetic energy,such as may be produced by civil and military aviation systems.

To deal with this problem, a few units have been designed with the chassisdynamometer mounted on a large turntable capable of being rotated through 360�.

Climatic test cells for vehicles

There is a requirement for this kind of facility for development work associatedwith various aspects of performance under extreme climatic conditions. Subjectsinclude driveability, cold starting, fuel waxing and solar loading, vapour lock, airconditioning and the ubiquitous topic of emissions.

The design of air conditioning plants for combustion air has been discussed inChapter 5. The design of an air conditioned chamber for the testing of completevehicles is a very much larger and more specialized problem. Figure 18.3 is aschematic drawing of such a chamber, built in the form of a recirculating wind tunnelso that the vehicle on test, mounted on a chassis dynamometer, can be subjected tooncoming air flows at realistic velocities, and in this particular case over a temperaturerange from +40 to −30�C.

The outstanding feature of such an installation is its very large thermal mass and,since it is usually necessary, from the nature of the tests to be performed, to varythe temperature of the vehicle and the air circulating in the tunnel over a wide range

378 Engine Testing

Control surface

Fan

Chassis dynamometer

Test vehicle

Fromheater

Cooler

Fromdehumidifier

Hv

HW

Hf

Hr

HW

Ha

Figure 18.3 Schematic of climatic wind tunnel facility fitted with a chassisdynamometer

during a prolonged test period. This thermal mass is of prime significance in sizingthe associated heating and cooling systems.

The thermodynamic system of the chamber is indicated in Fig. 18.3 and the energyinflows and outflows during a particular cold test, in accordance with the methodsdescribed in Chapter 2, were as follows:

The thermal capacity (thermal mass) of the various elements of the system isdefined in terms of the energy input required to raise the temperature of the elementby 1�C.

In . Out .Ha make-up air 20 kW Hr refrigerator circuit 495 kWHv vehicle convection/radiation 100 kW .Hr tunnel fan power 154 kW .Hw walls, etc. by difference 221 kW .Totals 495 kW . 495 kW

Ca air content of chamber, return duct, etc. volume approximately 550m3,approximate density 1.2 kg/m3, Cp = 1.01 kJ/kg.K, thermal mass =550× 1.2× 1.01= 670 kJ/deg C

Cv vehicle, assume a commercial vehicle, weight 3 tonnes, specific heat of steelapproximately 0.45 kJ/kg.K, thermal mass, say 3000× 0.45= 1350 kJ/deg C

Ce fan, cooler matrix, internal framing etc. estimated at 2480 kJ/deg C


Cs structure of chamber. This was determined by running a test with no vehiclein the chamber and a low fan speed. The rate of heat extraction by therefrigerant circuit was measured and a cooling curve plotted. Concrete is apoor conductor of heat and the coefficient of heat transfer from surfaces toair is low. Hence during the test a temperature difference between surfacesand wall built up but eventually stabilized; at this point the rate of coolinggave a true indication of effective wall heat capacity. The test showed thatthe equivalent thermal mass of the chamber was about 26 000 kJ/deg C,much larger than the other elements. Concrete has a specific heat of about8000 kJ/m3 deg C, indicating that the ‘effective’ volume involved was about26 000/8000= 3.2m3.

The total surface area of the chamber is about 300m2, suggesting that a surfacelayer of concrete about 1 cm thick effectively followed the air temperature.

Adding these elements together we arrive at a total thermal mass of30 500 kJ/deg C. Then since 1 kJ= 1 kW-sec, we can derive the rate at which thetemperature of the air in the chamber may be expected to fall for the present case.

This shows that the ‘surplus’ cooling capacity available for cooling the chamberand its contents, Hw, amounts to 221 kW. This could be expected to achieve a rateof cooling of 221/30 500= 0.0072 deg C/s, or 1�C in 2.3min. However, this is thefinal rate when the temperature difference between walls and air had built up to thesteady state value of about 20�C. The observed initial cooling rate is much faster,about 1.5�C/min. Heating presents less of a problem, since the heat released by thetest vehicle and the fan assist the process rather than opposing it.

A further effect is associated with the moisture content of the tunnel air. On awarm summer’s day this could amount to about 10 kg of water and during the coolingprocess this moisture would be deposited, mostly on the cooler fins, where it wouldeventually freeze, blocking the passages and reducing heat transfer. To deal with thisproblem, it is necessary to include a dehumidifier to supply dried air to the tunnelcircuit.

Special features of chassis dynamometers in climatic cells

Chassis dynamometers intended to operate over the temperature range +40 to −10�Ccall for no special features, apart from sensible precautions to deal with condensation;for temperatures below this range and particularly those working below −25�Ccertain special features will be required.

Until the compact ‘motor between rollers’ (‘motor-in-middle’ in USA) design ofdynamometer became common, the dynamometer motor in climatic cells was oftenisolated from the cold chamber and operated at normal temperatures. With the newdesigns two quite different strategies can be adopted to prevent low temperaturescausing temperature related variability in the dynamometer system.

380 Engine Testing

1500 Axle distance plus overhang Axle distance plus overhang 300 1000 – 2500 1550

Separate pipes containing powerand control cables

Pit ventilation pipes

Figure 18.4 A single roll set (2×2) chassis dynamometer plus road speedfollowing cooling fan, within a cell sized to test rear and front wheel drive cars(AVL Zöllner)

• The dynamometer pit can be kept at a constant temperature above or below thatof the chamber by passing treated (factory) air through the pit via ducts cast inthe floor (see Fig. 18.4). This strategy submits portions of the roll’s surface todifferential temperatures during the static cooldown phase but without reportedproblems.

• The pit can be allowed to chill down to the cell temperature but the parts ofthe chassis dynamometer that are crucial to accurate and consistent performance,such as bearings and load cell, are trace heated.

Whatever the layout components exposed to temperatures below 25�C such as rollsand shaft should be constructed of steel having adequate low-temperature strengthto avoid the risk of brittle fracture.

Independent wheel dynamometers

A limitation of the conventional rolling road dynamometer is that it is unable tosimulate cornering, during which the wheels rotate at differing speeds, wheel spinand skidding. With the increased adoption of electronically controlled traction andbraking (ABS), there is a growing requirement for test beds that can simulate theseconditions. There are two types of solution:

• Four roll-set rolling roads. These machines range from end-of-line rigs consistingof four sets of independent double-roller units, to complex development test rigs,some having steered articulation of each roll unit. The production rigs permit thechecking of onboard vehicle control systems, transducers and system wiring bysimulation of differential resistance and speeds of rotation. The development of


hybrid vehicles may give impetus to the development of independent four-wheelchassis dynamometers since some designs are based on individual wheel drivemotors. Electronically controlled transmission systems on 4× 4 vehicles alsorequire this type of dynamometer to simulate the road conditions for which theyare designed to compensate.

• Wheel substitution dynamometers. A major problem faced by engineers carryingout NVH testing on a chassis dynamometer is that tyre noise can dominatesound measurements. One answer is to absorb the power of each drive wheelwith an individual dynamometer. If the wheel hub is modified the vehicle canstill sit on its tyres, giving approximately the correct damping effect, while thetyre contact and windage noise is eliminated. Alternatively, the wheels may beremoved and the individual wheel dynamometers used to support the vehicle.The dynamometers should be four-quadrant machines, to simulate both drivingand coasting/braking conditions. Hydrostatic, a.c. and d.c. machines have beenused for this application.

The installation of chassis dynamometers

Whatever the type of unit the reader has to specify and install, there are some commonplanning and logistical problems that need to be taken into account. For single rollmachines built within a custom-built chamber, the longitudinal space requirementmay have to take account of front- and rear-wheel drive vehicles plus room for tie-down mechanisms and the vehicle cooling fan. A typical layout is shown in Fig. 18.4.

As with engine test cells dealing with the testing of exhaust emissions, chassisdynamometers associated with ultralow emissions work (SULEV) should be built onsites that are out of the drift zone of automotive and industrial pollutants.

Chassis dynamometer pit design and construction details

The minimum requirements, if not the exact specification, of the pit into which achassis dynamometer is to be installed will invariably be provided by the dynamome-ter’s manufacturer which will include the loading on the pit floor to aid the structuraldesigner. The pit must be built to a standard of accuracy rather higher than is usualin some civil engineering practice since all pit installed dynamometers have a closedimensional relationship with the building in which they are installed; this is par-ticularly true for the pit construction which has to meet three critical dimensionalstandards:

1. The depth of the pit in relation to the finished floor level of the test cell needsto be held to tight dimensional and level limits. Too deep is recoverable, tooshallow can be disastrous.

382 Engine Testing

2. The lip of the pit has to be finished with edging steel that interfaces accuratelywith the flooring plates that span the gap between the exposed rolls’ surface andthe building floor.

3. The centre line of the dynamometer has to be positioned and aligned with thebuilding datum and the vehicle hold-down structure that may also be cast intothe cell floor. Fine alignment can be achieved by the dynamometer installers,but movement at that stage will be very restricted.

Pit depth

To achieve point 1 above, it is normal practice to make the pit floor lower than thedatum dimension and then cast into the floor levelled steel ‘sole plates’ at the requiredheight minus 5–10mm to allow some upward adjustment of the machine to the exactfloor level. The final levelling can be done by some form of millwright’s levellingpads, or more usually levelling screws and shim plates, as shown in Fig. 18.5.

The system described above allows the pit floor to have some slight ‘fall’ towardsa small drainage sump fitted with a level alarm or float operated pump. If the sitehas a high water table, the pit will have been suitably ‘tanked’ before final concretecasting to prevent ground water seepage; however, spillage of vehicle liquids andcell washing will drain into the pit so a means of taking these liquids into a foulliquid intercept drain is advisable.

Steel sole plates laser levelled and epoxy grouted into pockets in concrete pit base

Levelling screws in each machine base to allow shim plates

to be inserted (backed off before HD bolts tightened)

Plastic (non-rusting) shim

plates

Holding down (HD) bolts

Concrete floor cured to full 21 day strength before installation

Figure 18.5 One of the levelled steel plates exactly levelled and cast in thepit floor to support the chassis dynamometer (one per dynamometer HD boltlocation)


The need to run subfloor cable and ventilation ducts, between the control roomand the pit and the drive cabinets and the pit requires preplanning and collaborationbetween the project engineer and builder, it can also complicate the task of creatinga watertight seal against ground water. Normally there will be the following subfloorducts:

• signal and separate small power cable ducts between control room and dynamome-ter pit;

• high power drive cables and separate signal cable ducts between the drive cabinetroom/space and the dynamometer pit;

• pit ventilation ducts which at a minimum will have to include a purge duct forremoval of hydrocarbon vapours (see ATEX requirements, Chapter 4).

On the wall of the test cell it is usual to fit a ‘break-out’ box that allows transducers,microphones or CAN bus plugs to be used for special vehicle-related communicationsbetween control room and test unit.

Project managers of any major test facility building are advised to check the layoutand take photographic records of all substructure service pipes and steel work beforeconcrete is poured.∗∗

When, due to diminishing access, the dynamometer has had to be located in itspit before the building has been completed and before permanent electrical servicesare available, it may be vulnerable until building completion to the pit flooding androll surface damage; provision must be made to guard against both eventualities.

Pit flooring

In most large chassis dynamometer projects, the false floor and its support structureare provided by the dynamometer manufacturer. It may be centrally supported by thedynamometer structure and will have access hatches for maintenance and calibrationequipment. The steel or aluminium floor plates will be lifted from time to time andany hatches will need to be interlocked with the control system. It will be clear toreaders that unless the pit shape is built to a quite precise shape and size, there willbe considerable difficulty in cutting floor plates to suit the interface and that errorsin this alignment will be highly visible. It is strongly recommended that the civilcontractor should be supplied with the pit edging material by the chassis dynamometermanufacturer and that it is cast into the pit walls under his supervision before finaledge grouting. In some cases the pit edging assembly can be fixed with jig beams intemporary place to hold the rectangular shape true while fixing (Fig. 18.6).

∗∗ AJM was once told by a steel fixer that his trade was comparable with that of a surgeon inthat they both bury their mistakes; henceforth detailed records have been made of prepourstructures.

384 Engine Testing

Floor place supported on vibration absorbing strip

Angle sections combined to form pit edging

Edging fixed and cast into

wall

Figure 18.6 Section through one form of pit edging required to support pitflooring

Design and installation of variable geometry (4×4)dynamometers

Many chassis dynamometers intended for testing four, or more, wheel drive vehiclesare designed so that the interaxle distance can be varied so as to accommodate arange of vehicles in a product family. The standard method adopted in 4× 4 designsis for one, single-axle dynamometer module to be fixed within the building pit andthe identical second module to be moved towards or away from the datum moduleon rails inside the pit. The range of interaxle distance for light duty vehicles (LDV)will be in the range of 2 to 3.5m, but for commercial vehicle designs the range ofmovement will be far greater. This range of movement is important in that it tendsto determine the design needed to accommodate that section of cell flooring that isrequired to move with the traversing set of rolls. In the LDV designs where the totalmodule movement is around 1.5m the moving plates can be based on a telescopicdesign with moving plates running over or under those fixed. Where the movementconsiderably exceeds 1.5m, then the moving floor designs may have to be basedon a slatted sectional floor that runs down into the pit at either end, while beingtensioned by counterweights or as part of a tensioned cable loop. All these designscan be problematic in operation and require good housekeeping and maintenancestandards.

The traversing module is normally moved by two or more electrically poweredlead screw mechanisms coupled to a linear position transducer so that a set positioncan be selected from the control room. For operational reasons, it is recommendedthat some positional indication, vehicle-specific or actual interaxle distance, is markedon the operating floor.


LDV units are usually traversed on flat machined surface rails having lubricatedfoot pads fitted with sweeper strips to prevent dirt entrainment; this is very similar tolong established machine tool practice and has to be installed with a similar degree ofaccuracy. Large commercial vehicle rigs often use crane traversing technology ratherthan that of machine tools with crane rails installed on cast ledges within the pit walldesign and the axle modules running within ‘crane beams’ running on wheels someof which are electrically powered.

Drive cabinet housing

The same precautions relating to the integration of high powered IBGT or thyris-tor controlled d.c. drives covered in Chapter 10 relate to those fitted to chassisdynamometers.

The drive and control cabinet for high power chassis dynamometers may beover 6m long, 2.2m high and, although only 600–800mm deep, will require spacefor front and sometimes rear access. These units are heavy with high centres ofgravity and vulnerable to damage through damp or dirty atmosphere; if positionedbefore the building is clean and heated then adequate provision for protection shouldbe made.

Offloading and positioning

Even small machines that are delivered in one unit require lifting equipment tooffload, transport and position them in the pit. Forty-eight-inch roll machines orlarger units may have to be positioned in already built or partly completed chambers;therefore, the installation will often require specialist equipment and contractors tolift and manoeuvre the machine in a building area of limited headroom and limitedfloor access as in the case of lined climatic chambers. It is a project phase thatis used as an example in the ‘Project timing chart’ section of Chapter 1, which isrecommended reading.

Tyre burst detectors

On all mileage accumulation rigs and others undertaking prolonged automatic testsequences, there should be some form of detector that can safely shut down the wholesystem in the event of a tyre deflating. The most common form is a limit switchmounted on a floor stand with a long probe running under the car at its midpointwith limit switches fitted one on each side. Any deflation will cause the vehicle bodyto drop, shutting down the test in a safe manner.

386 Engine Testing

Cable layout

The general treatment of this subject in Chapter 10 applies in the present case, butspecial attention should be given to the avoidance of electrical interference frompower cables with instrumentation lines. In many cases, all the cables running tothe chassis dynamometer will at some point run through plastic tubes set within theconcrete of the pit walls and plant room floors. These cable tubes should be allocatedto either power or signal and should have a drainage slight fall to the pit. Afterinstallation is complete, the tubes can be ‘stopped’ at both ends with acoustic andfire attenuating material.

Loading and emergency brakes

It must be possible to lock the rolls to permit loading and unloading of the vehicle.These consist of either disc brakes fitted to the roll shafts or brake pads applied tothe inside surface of one or both rollers. In normal operation these brakes have tobe of sufficient power to resist the torques associated with driving the vehicle onand off the rig. Pneumatic or hydraulically powered, they are controlled either bythe operational controls (vehicle loading) or by the safety instrumentation (EM stop)and are designed to be normally ‘on’ and require active switching to be off (machineoperational).

It is not considered good practice to rely solely on these brakes to bring the rigto rest in the case of an emergency. They should only be used to supplement thebraking effort of the main drive system. In addition there must be a well-thoughtout ‘lock-out’ system to ensure that all personnel are outside restricted areas beforestart-up, particularly important in large facilities where the operator may not be ableto see all areas.

Vehicle restraints

The test vehicle must be adequately restrained against fore and aft motion under thetractive forces generated, and against slewing or sideways movements caused by thewheels being set at an angle to the rig axis. It should be pointed out to inexperiencedreaders that while the vehicle has no momentum driving on a rolling road, the chassisdynamometer does have momentum and it operates in the opposite direction oftravel to that which the vehicle would have at equivalent velocity. Sudden brakingof the vehicle therefore will cause it to try to jump off the rolls backwards ratherthan forwards. Vehicle restraint is less of a problem when the powered wheels are


located between a pair of rolls rather than resting on a single roller; however, pairedrolls are only used for short duration test work because of their limited diameterand hence poor contact pattern. They are generally used for low power end-of-linetesting, where they are usually relied on as the only means of restraint so as to allowfor rapid loading and unloading of the test vehicles. Where full restraint systemsare used, the cell floor must be provided with strong anchorage points. There aremany different designs of vehicle restraint equipment, ranging from the use of largepneumatic bags to simple chain or loading strap attachments. Three different typesof restraint system may be distinguished:

• A vehicle with rear wheel drive is the easiest type to restrain. The front wheelscan be prevented from moving by fore and aft chocks linked by a tie bar. Therear end may be prevented from slewing by two high strength luggage straps withintegral tensioning devices; these should be arranged in a cross-over configurationwith the floor fixing points outboard and to the rear of the vehicle.

• Front wheel drive vehicles need careful restraint, since any movement of thesteering mechanism with the rolls running at speed can lead to violent yawing.Restraints may be similar to those described for rear wheel drives, but with thestraps at the front and chocks at the rear. The driver in manned runs shouldbe familiar with the handling characteristics of the vehicle in normal operation,while for unmanned operation the steering wheel should be locked; otherwise,disturbances such as a burst tyre could have serious results. Tyre pressures,usually set above the normal level, should be carefully equalized.

• Four-wheel drive vehicles usual rely on cross-over strapping at both ends, butdetails depend on fixing points built into the vehicle. Where strong tow pointsare conveniently supplied they can be used to connect via a short articulatedconnector to a floor fixed restraint pillar. When using straps, it is good practice totie the rear end first and then to drive the vehicle slowly ahead with the steeringwheel loosely held. In this way, the vehicle should find its natural position on therolls and can be restrained in this position, giving the minimum of tyre scrubbingand heating.

Guarding and safety

Primary operational safety of chassis dynamometer operators is achieved by restrict-ing human access to machinery while it is rotating. Access to the pit should behardwire interlocked with the EM stop system and latched EM stop buttons posi-tioned in plant areas remote from the control room, such as the drive room and cellextremities.

The small exposed segment of dynamometer rolls on which the tyres rotate is themost obvious hazard to operators and drivers when the rig is in motion and personnelguards must always be fitted. Such guards are a standard part of the dynamometersystem for good safety reasons, but perhaps also because it is the only operational

388 Engine Testing

piece of the manufacture’s equipment that is visible to the visitor after installationand usually will bear the company logo. Cover plates for the rolls for use when thevehicle is present are also advisable, particularly to prevent surface damage duringmaintenance periods.

Roll surface treatment

In the past, most twin roll and many single roll machines had no special treatmentapplied to the roll surfaces, which had a normal finish machined surface. Braketesters and some production rigs have a high-friction surface, which can give rise tosevere tyre damage if the machine is used incorrectly. The roll surface of modernsingle roll machines may be sprayed with a fine-grained tungsten carbide coating,but other common surfaces include emery grit or plain steel.

‘Road shells’

For development work, and particularly for NVH development, it is sometimesdesirable to have a simulated road surface attached to the rolls. Usually the requiredsurface is a simulation of quite coarse stoned asphalt; but a roughened roll surfacegreatly increases the windage noise of the rig itself.

Road shell design and manufacture is difficult. The shells should ideally be madeup of four or more segments of differing lengths and with junction gaps that areas small as possible and cut helically. It is rarely required to induce low frequencyvibrations in the test vehicle; if it is required, it should be generated by randomlydistributed variations in roll surface height to avoid possible resonances in the vehiclesuspension.

The most usual techniques for producing road surfaces for attachment to chassisdynamometer rolls are as follows:

• Detachable fibreglass road shells having an accurate moulding of a true roadsurface. These shells are usually thicker than the aluminium type, below, andin both cases the floor plates must be adjustable in order to accommodatethem.

• Detachable cast aluminium alloy road shells, made in segments that may bebolted to the outside diameter of the rolls. The surface usually consists of parallelsided pits that give an approximation to a road surface.

• A permanently fixed road surface made up of actual stones bonded into a rubberbelt which is bonded to the roll surface.

Most road shells are not capable of running at maximum rig speed because theyare difficult to fix and to balance, also they may not be capable of transmitting fullacceleration torque. It is therefore necessary to provide safety interlocks so that the


computerized speed and torque alarms are set to appropriate lower levels when shellsare in use. Road shells also change the effective rolling radius and the base inertiaof the dynamometer, requiring appropriate changes in the software constants.

Driver aids

The control room needs to be in touch with the driver in a test vehicle on a chassisdynamometer, either by voice transmission or visually by way of an aid screen whichmust be easily visible to him.

A major function of such a screen will be to give the driver instructions forcarrying out standardized test sequences, for example production test programmesor emissions test sequences. This kind of VDU display is often graphical in form,showing for example the speed demanded and the actual speed achieved. Sincein the case of emissions tests, the test profile must follow that prescribed withindefined limits, it is usual to include an error checking routine in the software to avoidwasted test time. The screen may also be used to transmit instructions to supplementinformation conveyed by the two-way voice link.

Fire suppression

For a more general treatment of the subject of fire suppression, see Chapter 4.The risk of a vehicle fire during chassis dynamometer running is quite high,

since the air cooling, even with supplementary fans, is likely to be less than thatexperienced on the road. Underfloor exhaust systems, in particular, can become veryhot and could ignite fuel vapour, whatever its source. A fixed fire suppression systemis more difficult to design than in an engine test cell because of both the large sizeof rolling road cells and the more difficult access to the seat of the fire, which maybe within the vehicle body.

All vehicle test facilities should be equipped with substantial handheld or hand-operated fire extinguishers and staff should be trained in their use. One methodof fire protection is to fit the test vehicle with a system of the type designed forrally cars which enables the driver or control room to flood the engine compartmentwith foam extinguishant. Automatic gas-based systems of the type used in someengine test cells are less effective in vehicle cells in view of the greater difficultyin ensuring that all personnel, including the driver, have been evacuated before theyare activated. The modern trend is towards water fog suppression systems, whichmay include discharge nozzles mounted beneath the vehicle and thus near the seatof most potential fires.

There must be a clear and unimpeded escape route for any test driver and theimpairment in vision from steam or smoke must be taken into account, particularly inthe case of anechoic cells, where the escape door positions should not be camouflagedwithin the coned surface.

390 Engine Testing

The effect of roll diameter on tyre contact conditions

The subject of tyre rolling resistance has given rise to an extensive literature, whichis well summarized in Ref. 1. The bulk of the rolling resistance is the consequence ofhysteresis losses in the material of the tyre and this gives rise directly to heating of thetyre. A widely accepted formulation describing the effect of the relative radii of tyreand roll is:

Fxr = Fx

(

1+r

R

)12

(2)

where:

Fxr = rolling resistance against drumFx = rolling resistance on flat roadr = radius of tyreR = radius of drum

Figure 18.7a shows the situation diagrammatically and Fig. 18.7b the correspondingrelation between the rolling resistances. This shows that rolling resistance (and hencehysteresis loss) increases linearly with the ratio r/R. For tyre and roller of equal

r

r /R

R

(a)

(b)

2

1

00 1 2 3 4

µ

µ 0

Figure 18.7 (a) Relationship between tyre and chassis dynamometer rollerdiameters and (b) linear relationship between rolling resistance and the tyre/rollratio


diameter the rolling resistance is 1.414× resistance on a flat road, while for a tyrethree times the roller diameter, easily possible on a brake tester, the resistance isdoubled. A typical value for the ‘coefficient of friction’ (rolling resistance/load)would be 1 per cent for a flat road.

To indicate magnitudes involved, a tyre bearing a load of 300 kg running at17m/s (40mph) could be expected to experience a heating load of about 500W ona flat road, increased to 1 kW when running on a roller of one third its diameter.Clearly the heating effects associated with small diameter rolls are by no meansnegligible.

Tyres used even for a short time on rolling roads may be damaged by heating anddistortion effects, which are aggravated by the fact that the roll will also be heatedin the course of the run. Some dynamometer systems are fitted with tyre coolingsystems to reduce tyre damage but all tyres used for any but short duration testsshould be specially marked and changed before the vehicle is allowed on publicroads.

Specification, accuracy and calibration

A typical specification for a chassis dynamometer will include the following infor-mation as a minimum:

• base inertia;†

• minimum simulated inertia;• maximum simulated inertia;• continuous rated power, absorbing mode;• continuous rated power, motoring mode;• torque range, absorbing and motoring, sometimes in the form of tractive effort at

base speed;• number of rolls;• roll diameter with tolerance (typically <±0.005 inch on a 48-inch roll);• roll concentricity (typically <±0.1mm (0.004 inch) on a 48 inch roll);• maximum surface speed (km/hour);• maximum permitted acceleration;• maximum time to stop from maximum rated speed under regenerative braking;• type of drive protection system (supply failure during full regenerative load, etc.).

These parameters determine the basic geometry of the rig and the size of motor andits associated variable speed drive system.

† It is common practice to use the term ‘inertia’ to mean the simulated or equivalent vehicleinertia, rather than the moment of inertia of the rotating components. The inertia is thusquoted in kg rather than kgm2.

392 Engine Testing

Where a purpose of the dynamometer is to carry out homologation tests, theseshould be clearly specified, together with the range of vehicles to be covered; theacceptance procedure should include performance of these tests.

The essential measurements performed by a chassis dynamometer, equivalent tothe measurement of torque and speed in a conventional machine, are roller surfacespeed and the force (the tractive effort at the roll periphery) transmitted to or fromthe rolls.

This force is usually not measured directly, but derived from a torque measurementat the motor, performed either in the conventional manner by utilizing a trunnion-mounted motor with torque arm and load cell, or by some form of torque shaftdynamometer. Whatever method is used, there will be an in-built error arising frombearing losses and roll windage: these losses can be measured by running the machinethrough the speed range with no vehicle present and noting the necessary drivingtorque, which should be added to the observed torque when absorbing power andsubtracted when braking, to give the true tractive force at the roller rim. (This is notthe complete answer as there will be a further, unquantified, loss due to roll bearingfriction when the rolls are under load.)

Some dynamometers include these corrections in the stated dynamometer con-stants (see Chapter 8 for a detailed discussion of torque calibration procedures). Thisis often not a simple matter in the case of chassis dynamometers as the torquesinvolved can be very substantial and space very restricted: the preferred methodusing graduated weights to apply torque in 10 steps may not be practicable anddynamometers are often supplied with a single mass corresponding to full torqueplus a few small weights adding up to about 10 per cent of full scale torque.

The accuracy of calibration demanded and the calibration methods to be used maybe included in emissions legislation relevant to chassis dynamometer testing as partof the general tightening of standards.

Equivalent road speed measurement will usually be by means of a high resolutionoptical encoder on the roll shaft, thus introducing a possible source of error arisingfrom uncertainty regarding true roll diameter, particularly the case with coated rolls.It is common practice to use a so-called ‘pi-tape’ to obtain the effective diameter bymeasuring the circumference.

A check on the actual base inertia of the dynamometer is made either by means ofa number of ‘coast-down’ tests or by both accelerating and decelerating the rolls atan approximately constant rate, plotting motor torque and speed in both cases. Thelatter method is the more accurate since it effectively eliminates the internal drag ofthe dynamometer motor.

It will be apparent that really accurate measurements of traction/braking forceare by no means easy to achieve. Much depends on the quality of manufacture andconsistency of performance of the machine, the accuracy of the basic measurementsof torque, roller diameter and speed and the capability of the associated computer.

A common operational problem experienced in test laboratories is that after rigginga vehicle on the chassis dynamometer and taking considerable time and effort toconnect all of the required emission equipment and transducers, it is then required


to perform warm-up and coast-down checks on the dynamometer, during which thevehicle should be removed. The solution is provided on most high end designs bythe incorporation of a mechanism that arcs around the drum and lifts the vehicle tyreon two small rollers just clear of the drum surface. On completion of the warm-uproutine, the mechanism arcs down, out of contact, and the tyre is returned to itsoriginal position.

Limitations of the chassis dynamometer

Like all devices intended to simulate a phenomenon, the chassis dynamometer hasits limitations, which are easily overlooked. It cannot be too strongly emphasizedthat a vehicle, restrained by elastic ties and delivering or receiving power throughcontact with a rotating drum, is not dynamically identical with the same vehicle inits normal state as a free body traversing a fixed surface. During motion at constantspeed, the differences are minimized, and largely arise from the absence of air flowand the limitation of simulated motion to a straight line.

Once acceleration and braking are involved, however, the vehicle motions in thetwo states are quite different. To give a simple example, a vehicle on the road issubjected to braking forces on all wheels, whether driven or not, and these give riseto a couple about the centre of gravity, which causes a transfer of load from the rearwheels to the front with consequent pitching of the body. On a chassis dynamometer,however, the braking force is applied only to the driven wheels, while the forwarddeceleration force acting through the centre of gravity is absent, and replaced bytension forces in the front-end vehicle restraints (remember: braking on a rollingroad tends to throw the vehicle backwards).

It will be clear that the pattern of forces acting on the vehicle, and its consequentmotions, are quite different in the two cases. These differences make it very dif-ficult to investigate vehicle ride and some aspects of NVH testing, on the chassisdynamometer. A particular area in which the simulation differs most fundamentallyfrom the ‘real-life’ situation concerns all aspects of driveline oscillation, with itsassociated judder or ‘shuffle’.

The authors have had occasion to study this problem in some detail and, whiletheir analysis is too extensive to be repeated here, one or two of their conclusionsmay be of interest:

• The commonly adopted arrangement, whereby the roll inertia is made equal tothat of the vehicle, gives a response to such disturbances as are induced bydriveline oscillations, i.e. judder, that are significantly removed from on-roadbehaviour.

• For reasonably accurate simulation of phenomena such as judder, roll inertiashould be at least five times vehicle inertia.

• Electronically simulated inertia is not effective in this instance: actual mass isnecessary.

394 Engine Testing

• The test vehicle should be anchored as lightly and flexibly as possible: this is notan insignificant requirement, since it is desirable that the natural frequency of thevehicle on the restraint should be at the lower end of the range of frequencies,typically 5–10Hz, that are of interest.

It is strongly recommended that investigations of vehicle behaviour under conditionsof driveline oscillation should proceed with caution if it is intended to involve runningon a chassis dynamometer.

Summary

The subject of rolling roads and chassis dynamometers is a very extensive one,worthy of a textbook in its own right. The application of these machines, at all butan elementary level, is very largely in the field of transient testing and the road loadequation. However, the testing techniques employed have a great deal in commonwith those applicable to engine testing dynamometers, and almost the whole of thepresent book is relevant.

In this chapter, an attempt has been made to survey the whole range of apparatusthat may be described by the general heading of rolling roads/chassis dynamometers,from the simple garage brake tester to complex and very costly climatic chambersand anechoic cells. The design of the associated test cells, vehicle restraint systems,driver communications and safety precautions have been discussed and problems ofaccuracy and calibration considered.

The subject of the building/dynamometer interface has been covered in somedetail because errors discovered at installation are expensive to rectify.

Reference

1. Tire Rolling Resistance, Symposium, 122nd Meeting, Rubber Division, American ChemicalSociety, 1982.

19 Data collection, handling,post-test processing, enginecalibration and mapping

Introduction and definition of terms

One of the primary purposes, and in most cases the only purpose, of an engine orvehicle test cell is to produce data. The collection, verification, manipulation, display,storage and transmission of these data should be amongst the prime considerationsin the design and operation of any test facility.

It is perhaps the speed of recording, processing, storage of data and the creationof engine ‘models’ that has been responsible for the most revolutionary changes inthe practice of engine testing in the last 10 years.

Test techniques now available are only made possible by the ability of moderncomputerized systems to create ‘virtual engine’ models. These models are an essentialprerequisite for that part of the engine development process known as calibration.

The process of mapping is part of the process by which engine calibration modelsare initiated and then developed. In this context, the process refers to the processof creating and testing the three-dimensional control ‘maps’ held within the enginecontrol unit that determines not only ignition timing and fuel injection at all statesof engine speed and torque, but several other engine and vehicle control parameters.

Computerized record systems are no more inherently accurate, nor traceable, thanpaper systems. All data should have an ‘audit trail’ back to the calibration standardsto prove that the information is accurate to the level required by the user. It is a primeresponsibility of the test cell manager to ensure that the installation, calibration, record-ing and post-processing methods associated with the various transducers and instru-ments are in accordance with the maker’s instructions and the user’s requirements.

Bad data are still bad data, however rapidly recorded and skilfully presented.

The traditional approach of test data collection

It should not be assumed that traditional methods of data collection should necessarilybe rejected, they may still represent the best and most cost-effective solution incertain cases where budgets and facilities are very limited.

Figure 19.1 shows a traditional ‘test sheet’. The sheet records the necessaryinformation for plotting curves of specific fuel consumption, b.m.e.p., power output

DATE

DYNAMOMETER TORQUE ARM

ENGINE VARIABLE COMP

BAROMETER 755 mmHg

COUNTERTACHO

rev/min rev

2x

2000 2012

2002

1022

2025

1989

2010

2019

1995 74.5

74.5

74.5

74.0

71.0

68.0

67.0

50.5 2.022

3.726

3.630

3.993

4.089

4.160

4.178

4.129

3.601

3.347

2.903

1.944 249

370

427

464

530

491

530

530

527

506

484

477

47.0 3.830 1.357

0.966

0.945

0.791

0.699

0.627

0.556

0.535

0.502

0.513

0.530

0.580

0.809 682

673

685

675

500

620

ho

TA

Va

ma

6.15

6.15

6.13

6.10

6.05

6.00

5.97

5.93

5.93

5.87

5.95

6.03

6.03 19

19

21.5

21 4.02 0.827 6.89

5.89

5.88

5.87

5.84

5.82

5.80

5.79

5.79

5.76

5.80

5.84

5.84 17.83

16.66

15.70

14.98

14.20

12.58

11.99

10.71

9.81

8.92

8.23

7.85

7.38

0.630

0.655

0.620

0.620

0.616

0.621

0.609

0.620

0.618

0.623

0.626

0.629

4.92

4.91

4.90

4.88

4.86

4.85

4.83

4.83

4.81

4.84

4.88

4.88

3.600

3.429

3.158

2.857

2.609

2.323

2.209

1.957

1.846

1.773

1.682

1.572

50.0

52.5

57.0

63.0

69.0

81.5

92.0

97.5

101.5

107.0

114.5

77.5

360

3.9002035 69.0

1988 65.2

2008 60.0

2010 52.0

2000 35.0

780

834

891

962

1044

1156

1304

1355

1560

1615

1698

1792

1908

DENSITY OF FUEL 0.75 kg/litre

WEAK

RICH

n kW kN/m2FN

sec

rev/min

BRAKELOAD

POWER b.m.e.p1sec50/cc litre/

hour

litre/kW hr

EXH°C HEAD

cmH2O

TEMP°C

VOLF/R I/s

EFFY.η VOL

MASSF/R: g/s

a/IRATIO

AIR CONSUMPTION MEASUREMENTFUEL

CUSTOMER

265 mm

85 mm

b.m.e.p =POWER kW =

FNn

36,0007123 F

N kN/m2

AIR TEMP. 21°C

STROKE 825 mm

AIR BOX SIZE

FUEL Litre/Hour =

CYLINDERS 1

150 litre

SWEPT VOL. 468 cc FUELSHELL

4 -STAR

ORIFICE DIA. 18.14 mm FLOWMETER NO.

NOTE:

BORE

TEST SHEET– I.C. ENGINES UNIT NO.

WORKS ORDER NO.

REMARKS

COMPRESSION RATIO 7.1

FUEL GAUGE ‘O’

OILSHELL20W 30

1801

Figure 19.1 A traditional test sheet

Data collection, handling, post-test processing, engine calibration 397

and air/fuel ratio of a gasoline engine at full throttle and constant speed, with a total of13 test points. All the necessary information to permit a complete and unambiguousunderstanding of the test, long after the test engineer has forgotten all about it, isincluded. The same cannot necessarily be said of computer-acquired test data thathas been poorly integrated into a data management system. It is a great temptationto leave the computer to ‘get on with the job’, without paying sufficient attention tothe proper annotation of the information it has generated.

Such simple test procedures as the production of a power–speed curve for a rebuiltdiesel engine can perfectly well be carried out with recording of the test data by handon a pro-forma test sheet. In spite of the low cost of computer hardware there maybe little justification for the development and installation of a computerized controland data acquisition system in the engine rebuild shop of, for example, a smallcity bus company handling perhaps 20 engines per month. If an impressive ‘birthcertificate’ for the rebuilt engine is required, the test sheet may easily be transferredto a computer spreadsheet and accompanied by a computer-generated performancecurve.

In the data flow shown in Fig. 19.2, the only post-processing of the test data is theconversion into a delivered engine ‘birth certificate’ which may be simply a reducedand tidy copy of the original data. The archive is available for use for technical andcommercial statistical analysis.

Data archive

Engine test

Engine birthcertificate

Test results sheet

Data archive

Figure 19.2 Data flow based on pre-computerized systems

398 Engine Testing

Chart recorder format of data

Chart recorders represent a primitive stage in the process of automated data acqui-sition. Multipen recorders, having up to 12 channels giving continuous analoguerecords in various colours, still have a useful role, particularly in the testing of dis-persed systems, such as installed marine installations or of large engines ‘in the field’where robustness and the ease with which the record may be annotated are of value.

It is the manner in which the data are displayed as a coloured graph of valuesagainst time, now emulated by computerized displays, which is the real value to testengineers who need the clear and immediate indication of trends and the interactionof data channels.

Computerized data recording with manual control

As the number of channels of information increases beyond about 12 and there is aneed to calculate and record derived quantities, such as power and fuel consumption,the task of recording data by hand becomes tedious and prone to error.

An operational example, at a slightly higher level of complexity than the previousone, might be a small independent company producing specially tuned versions ofmass-produced engines for sports car use. The test cell will be required to producedetailed power curves and will also be used for development work.

Information to be recorded is likely to include

• torque, speed and fuel consumption;• fluid temperatures and pressures;• several critical gas and component temperatures.

The total required number of data channels could exceed 40. Typical equipmentwould be a PC-based data acquisition package with 16 pressure channels, 16 tem-perature channels, four channels for receiving pulse signals plus torque transducersignal processing and a facility for processing input from a fuel consumption meter.

In this test scenario, no control, either of the test sequence, of the dynamometeror the engine itself, is exercised by the PC; these are under manual control of theoperator. A ‘snapshot’ record of all data channels may be taken and written to discevery time the operator presses a particular key of the test cell keyboard. Each suchrecord will be time and date stamped with a ‘time of test’ figure derived from thePC’s internal clock. As a simple post-processing exercise, the operator is able to callup a power curve or any other preprogrammed graph from any of these data setsusing spreadsheet software.

It is entirely possible to adapt a PC data acquisition system at this level to giveautomatic test sequence control without large additional expenditure, provided thecell is equipped with dynamometer and engine controllers designed to accept externalcommands.


However, if there is a wide variety of test sequences, not repeated very often, itcan take longer to modify or programme the test sequence than to run the tests undermanual control.

At all levels of complexity, the appropriate degree of sophistication in the man-agement of the test and the recording of the data calls for careful consideration.

Analogue to digital (ADC) conversion accuracy

Signals from transducers having analogue outputs are digitized, conditioned andlinearized in the PC for display and stored in the chosen style and form. Proprietarydata collection and signal conditioning cards are commonly available with ADC of8-, 12- and 16-bit resolution. These should be selected as appropriate to the accuracyand the significance of the signals to be processed.

It is entirely inappropriate and gives a false illusion of accuracy (see Chapter 20) touse 16-bit or higher resolution of ADC for measurement channels having an inherentinaccuracy an order of magnitude greater. For example, a thermocouple measuringexhaust gas temperature may have a range of 0–1000�C and an accuracy of >±3�C.The 16-bit resolution would be to about 0.016�C, while the entirely adequate 12-bitresolution would be 0.25�C.

The modern standard: full computerized control and dataacquisition

Fully computerized test cells are inevitably computer dependent. If the computer andits software are not functioning, then the test bed may be effectively unusable. Thedegree of dependency is a function of the detailed design of the equipment; somedesigns will allow manual operation with the computer out of action but many willnot. In almost all systems a pseudomanual control mode is available, whereby theoperator can start and stop the engine, then change speed and torque via a manuallyoperated panel; however, this mode is invariably fly-by-wire and computer-dependentfor all but demand input.

Certain tests, such as those required for engine emission homologation, may bepreprogrammed in all details by the control system supplier, but normally the user isprovided with a set of software tools to set up the required functions: data collection,labelling, scan rates, logging frequency and display of data. It is vital that trainingin the efficient use of these tools is given to those in the whole chain of data use.

There should be the ability to set a limited number of channels to higher scanrates than others, but it is possible to place too much emphasis on scan rates whencomparing systems. In certain types of transient test logging rates of 1 kHz or moremay be required, but for many parameters, fluid temperatures for example, such scanrates are inappropriate and may even, as in the case of engine speed, give misleading‘snapshots’ of the value being measured.

400 Engine Testing

Higher than necessary scan rates and greater than necessary levels of analogueto digital conversion can cause unnecessary problems in processing speed and datastorage.

‘Real-time’ display of selected data, often in a graphical format chosen by theoperator, is becoming more common as the power of the test bed computer increases;such displays can be very useful when setting up control loops. However, the prolif-eration of screens showing test data brings with it unique problems of ‘informationoverload’ for the operator and some degree of convention and management is requiredto discipline the use of upward of six screens of information now commonly seen inR&D cells.

Data collection and processing in engine mappingand calibration

In this usage, the term ‘calibration’ is used to describe the process by which theoperational characteristics of one physical engine (more correctly, power-train unit)is changed, through manipulation of all its control mechanisms to suit a targetgeographical/vehicle/driver combination while remaining within legislative emissionlimits.

It costs a vast amount of money to develop a new engine and the enterprise is onlyeconomically viable if the engine then finds the widest possible field of application.Today this field is essentially worldwide and the developers will hope that theirengine will be incorporated in a very wide range of vehicles, manufactured in manydifferent countries, each having their own customer preferences, driving habits andlegislative requirements.

The optimized control variables, when found, are stored within the ‘map’ heldin the engine/power train, engine management unit or engine control unit (ECU)and in the vehicle control systems, such as ABS braking and stability control. Thus‘mapping’ an engine, outside its intended vehicle, requires that it be run on a testbed capable of simulating not only the vehicle’s road load characteristics but also thefull range of the vehicle’s performance envelope, including the dynamic transitionsbetween states.

In order to reduce the testing time required to produce viable ECU maps, andin order to lower the complexity of the operator’s role in the mapping process, theengine test industry is continuing to develop powerful software tools designed tosemi-automate the tasks.

During the mapping process, in order to discover boundary conditions, such as‘knock’, the engine will be run through test sequences that seek, and then confirm,the map position of these boundaries at a wide range of driver operating conditions.Conditions of excessive temperatures, cylinder pressure and ‘knock’ require fastdetection and correct remedial action. Mapping tools such as CAMEO®, produced byAVL, contain test sequence control algorithms which allow the automatic locationof these boundary conditions and can react correctly when they are located.


A typical mapping sequence is that called ‘spark sweep’ where at a given operatingpoint the spark timing is varied until a knock condition is approached, the timing isthen backed off and the knock point approached with finer increments to confirmits map position. The test sequence then sweeps the ignition timing so that a faultcondition, such as high exhaust temperature, is located at the other end of the sweep.Given sufficient data, the software can then calculate and extrapolate the boundaryconditions for a wide operating envelope without having to run all of it.

Advanced design of experiment (DOE) tools, such as those embedded in productslike CAMEO, allow the test engineer to check the plausibility of data, the generationof data models and obtain visualization of system behaviour while the test is inprogress.

The ultimate target of the calibration engineer is to be able to set the operatingprofile of the test engine, then set the sequence running, only to return when thetask of producing a viable engine map has been automatically completed; this willbe achieved within the lifetime of this edition.

Use of data models in ‘hardware-in-loop’ testing

Once a viable model of an engine exists it can be used in the testing of engine andvehicle subsystems. Of particular value is the ability to inject fault conditions into thewhole-vehicle safety matrix without physically endangering it, or the test engineer.

The various test techniques use the concept of a loop of interactivity in whichinputs are injected into either an actual module or a virtual module and the outputsused in a feedback loop. Common examples are known by the following acronyms:

• HIL, hardware in the loop, which allows the experimenter to test a real componentin a virtual environment in closed-loop mode. Typically the real component is anengine electronic control unit (ECU).

• SIL, software in the loop, where software components (e.g. SW of an ECUcontroller) can run on a test platform.

• MIL, model-in-the-loop, where control models can be run in a virtual environment(e.g. hardware such as the ECU is simulated).

Clearly, these test techniques are only as good as the models on which they arebased which puts heavy responsibility on the quality control and management ofdata. However, as the total store of verified data builds, the technique improves andis able to shorten development times in the calibration of vehicle control systemsand engine peripherals.

In-vehicle data transfer protocols

The electronically controlled engines of today are installed within electronically con-trolled vehicles; this has led to the development of various designs of in-vehicle data

402 Engine Testing

bus. The CAN bus (controller area network) has come to dominate the automotiveindustry particularly in Europe. To ensure that CAN bus devices from different sup-pliers interoperate and that a good standard of EMI and RFI tolerance is maintained,the SAE J1939 family of standards has been developed. Pressure on the data transferrates beyond the capabilities of the original CAN bus have meant that several otherprotocols and standards (some optical based) have been developed and it is not clearwhich will become a future standard. The test engineer dealing with engine ECU andprimary vehicle systems such as emission control or transmission control will needto keep up to date with both J1939 systems and any sector specific standards suchas SAE J1587/1708 which covers heavy truck systems; these systems will appear inthe connection of modules during HIL testing system.

Smart devices and systems: communication and control

The main test bed computer may be linked to smart instrumentation, sending thenecessary triggering or control signals and recording the formatted data produced bythe device. Linking the test bed control computer with the many types of instrumentavailable is not a trivial task and it is essential, when choosing a main controlsystem, to check on the availability of device drivers and compatibility of third partyequipment and data transfer protocols.

Software for production test cells

Computers required to manage production test cells are required to perform a numberof specialized tasks, of which examples follow.

Bar code labels or embedded read/write chips are commonly attached to enginesarriving for test; these will contain all the information necessary for the test standsoftware to select the required test procedure with appropriate values, and to createa test ‘header’ that will form the basis of the engine test certificate. Following thetest, the test bed computer may print a bar code label or write test information to theengine chip, which may include fault codes for use in the minor repair stations.

The computer software for production test beds must contain much more informa-tion about the end user and the production logic of the plant than would be embodiedin the software for an R&D test facility. The identity codes for engines, the procedurefor dealing with minor repairs, the pass and fail criteria, integration with productioncontrol computers and many other details must be built into the application softwareat the time of facility design.

Production test cells often require the operator to make visual checks and perhapsmake some physical adjustment, such as setting an idling stop, in the course of anotherwise automatic test sequence. An operator may supervise several cells and sowill need to be called to a particular unit when his intervention is required, either for


normal or for abnormal events. Flashing signals and clear display messages callingfor acknowledgement will be required.

There are legal requirements in most countries concerning the archiving of prod-uct data

Data processing in cold testing

The vibration signature of rotating equipment has long been used as a diagnostic toolby engineers, particularly in the industrial transmission industry where noise analysisat discrete frequencies allowed tooth contact anomalies to be identified before anyphysical evidence could be detected.

The availability of ever improved transducers and high-speed data processinghas meant that engine developers can now study the vibration patterns produced byengine components forming part of a fully built engine.

A key element of cold testing is the development and maintenance of vast amountsof data relating to the vibration characteristics of a ‘good’ engine so that the charac-teristics of a ‘bad’ engine may be recognized by a computerized system using patternrecognition technology.

While gross build errors, such as valve timing or missing components, are easyto identify, problems with individual components such as fuel injectors will have amore subtle signature and the limits of patterns that are within an acceptable bandhave to be set empirically.

Typically, cold test is carried out on automotive gasoline and diesel engines whereit offers manufacturers considerable advantages over the traditional hot test becauseit can take a quarter of the time and be carried out within a cheaper infrastructure.When the same vibration signature evidence is sought from a hot test, ‘the bang’gets in the way. Cold testing commonly includes

• signature of oil pressure over time to check oil pump and oil circuit integrity;• ‘torque to turn’ of the engine over the test sequence gives indication of tight

pistons or bearing shells;• crank and cam timing;• fuel rail pressure pulsations checking fuel supply and injectors;• inlet and exhaust airflow and compression check valve operation.

Electrical harness integrity testing may also be carried out at the same test station.However, there are some shortcomings inherent in the cold test process, two of

which are that leaks induced by differential component expansion during the heatingof the engine are not evident and the fluid flows are lower than hot test so thatmanufacturing debris may not be flushed.

It should be noted that in the case of cold testing of diesel engines the fuelinjection timing has to be adjusted to prevent the engine running but still allowingfuel to flow, albeit for only a few pulses during the test.

404 Engine Testing

A cold test station does not require much of the test facility infrastructure of a hottest stand. Dedicated ventilation and exhaust systems are absent, hazard containment,fuel system and noise attenuation are simplified, fire risks are lower. The engine ismotored by an a.c. dynamometer.

It is the acquired data of the engine’s whole life and the application of those datainto the recognition of incipient faults in the engine at the time of testing that is themost important asset to the engine builder; such application requires dedicated staffand a policy of continuous improvement.

Anechoic testing and model creation

Modern digital, sound recording, devices and modelling software have reduced theamount of engine or vehicle running in anechoic cells almost to the minimum requiredto create a sound generation model over the operating range or at the point ofinterest and then test the results of modifications made based on the analysis of thosemodels.

Human intervention may still be required in order to judge the level of distressor annoyance induced by noises of the various sound patterns recorded. This, as anyparent of teenage children knows, is a highly subjective subject. The more difficultpart of the development process is to identify the physical modifications to thepowertrain system required to change the noise to the acceptable profile that hasbeen produced by manipulation of the original digital recording.

Management of data: some general principles

It cannot be too strongly emphasized that present-day data acquisition systems arecapable of storing vast amounts of information that may be totally irrelevant to thepurposes of the test or stored in a way that makes it virtually impossible to correlatewith other tests.

The first basic rule is to keep a consistent naming scheme for all test channelnames, without which it will not be possible to carry out post-processing or com-parison of multiple tests. Since most engine and vehicle test work is relative, ratherthan absolute, this ability of comparison and statistical analysis of results is vital.

Mere acquisition of the data is the ‘easy’ part of the operation. The real skill isrequired in the post-processing of the information, the distillation of the significantresults and the presentation of these at the right time to the right people in a formthey will understand. This calls for an adequate management system for acceptance(rejection), archiving, statistical analysis and presentation of the information.

Data are as valuable as their accessibility to the people that need to use them.The IT manager, in charge of the storage and distribution network of the data, isimportant in the test facility team and should be local, accessible and involved. One


vital part of the management job is to devise and administer a disaster recoveryprocedure and back-up strategy.

Normally the IT manager will have under his direct control the host computersystem into which, and out of which, all the test bed level and post-processingstations are networked. Host computers can play an important role in a test facility,providing they are properly integrated. The relationship between the host computer,the subordinate computers in post-processing offices and the test cells must becarefully planned and disciplined. The division of work between cell computerand host may vary widely. Individual cells should be able to function and storedata locally if the host computer is ‘off-line’, but the amount of post-processingundertaken at test bed level varies according to the organizational structures and staffprofiles.

In production testing, the host may play a direct supervisory role, in which caseit should be able to repeat on its screen the display from any of the cells connectedto it and to display a summary of the status of all cells.

With modern communications it is possible for the host computer to be geograph-ically remote; some suppliers will offer modem linking as part of a ‘site support’contract. This ability to communicate with a remote computer raises problems ofsecurity and safety since it may be possible, and allowed, for remote staff to causemachinery to operate.

Post-acquisition data processing and reporting

Post-processing and display of data is often carried out using spreadsheet or relationaldatabase and graphical software that can read the buffered or stored informationdirectly from the test cell or host computer. It may be important for purchasers of dataacquisition systems to check that the data from their system can be converted into anon-specific format, such as ASCII, for importing into their preferred spreadsheetsand databases.

A tool widely used by high-end research laboratories is a software suite producedby AVL List called Concerto®, which is specifically designed to deal with thedifferent types of data from test bed, combustion analysis and emission equipmentand provides most of the mathematical, statistical and graphical tools required byautomotive development engineers. One advantage of such a tool is that security ofdata is built into the code so that accidental or deliberate changes to data can be seenor prevented.

If the ‘customers’ of the test facility are not aware of the capabilities of the dataacquisition system they will not make optimum use of it. It is essential that theyshould be given proper training in the capabilities of the facility and its data in orderto use it to maximum advantage. This statement will appear to be a truism yet, in theexperience of the author, training in the capabilities of modern test equipment anddata analysis is often underfunded by purchasers after the initial investment, whichleads to skills and use levelling out below optimum.

406 Engine Testing

Physical security of data

The control room of an engine test cell is rarely the ideal environment for storingtest data. Rules are required for the management and storage of data, whether locallyor at the host computer and access to the various functions must be restricted if theintegrity of data and system control is to be preserved.

In most large engine test facilities there are three levels of password enforcedsecurity:

1. Operator. Typically at this level tests may be started, stopped and run onlywhen the ‘header information’, which includes engine number, operator nameand other key information, has been entered. Operator code may be verified.No alarms can be altered and no test schedules edited, but there will be a text‘notepad’ for operator comments, to be stored with the other data.

2. Supervisor. On entry of password this level will allow alteration of alarm levelsand test schedules, also activation of calibration routines, in each case by wayof a form fill routine and menu entry. It will also give access to fault findingroutines, such as the display of signal state tables.

3. Engineer. On entry of the password, this level is allowed access to debuggingroutines and language level editing.

Some test facilities, such as those involved with military or motor-sport developmentcan be under constant and real threat to the confidentiality of its data. Physicalrestriction of access of personnel is built into the design of such test facilities, butmodern miniaturized data storage devices and phone cameras seem to conspire tomake security of data more and more difficult; it is a management role to balancethe threat with sensible working practices.

A note on passwords

Passwords are keys that can be abused and misused. Military practice requires thatpasswords should be changed at random intervals and this is a sound principle. Such arequirement can be computer generated by the IT management. Increasingly, securitycard access to secure areas and computers is being used in place of passwords. Thisall tends to bring with it greater monitoring of the data handling activities of the cardholder.

General guidelines for the choice of engine test software

Worldwide, there are probably fewer than 15 proven suites of software that havebeen developed, and are supported, primarily for the control of R&D engine testbeds. Developers requiring the creation and manipulation of ‘virtual engine’ modelshave an even more limited choice.


The available proprietary suites of test cell control and data acquisition software,together with the range of engine test hardware they support, can be roughly dividedinto the same market categories as the users:

1. after-market, engine tuners and engine rebuilders;2. tier 2 automotive parts suppliers, undertaking endurance testing or quality assur-

ance testing and university departments;3. OEM, research institutes and international motor sport teams.

The stratification of price, complexity and adaptability matches these categories.The established suppliers of engine test cell computer systems will be familiar

with all the usual safety logic applicable to test cell control and will have included itin the design of their equipment and software. Therefore the prospective user needsto investigate what is successfully used in his peer group and to define his strategyfor data management within his operational specification (Chapter 1) before decidingon a supplier shortlist.

The screen displays should be assessed for the ease with which they may beadapted to the user’s requirements, their relevance to the tasks to be undertaken,and the clarity of the logic underlying the menu and display hierarchy. It is tobe remembered that the almost infinite variety of display that is possible must bemanaged and used in a safe and auditable manner. Errors and misuse do not merelycorrupt data, they can also lead to dangerous failures.

It is important to check that the control equipment can receive and transmit signalsto and from associated test equipment of third party suppliers. Special software driversmay have to be written and this work can be complex and expensive. Interfacing withsuch equipment as exhaust emissions analysers can be very difficult if not plannedfor when the system is designed.

Finally, the choice of equipment must be reviewed in the light of companypolicy and the medium-term development strategy for computerization. In particular,consider the extent to which your choice will tie you to a specific software or hardwareplatform and whether those platforms are sufficiently open or widely available andsupported by companies other than the supplier. It is inevitable that modificationsto the software in the light of experience will be necessary and the cost of softwaresupport and training should be factored into the purchase price (cost of ownership).

Further reading

SAE J1939 standards collection on the web at: http://www.sae.org/products/

j1939a.htm.

20 The pursuit and definition ofaccuracy: statistical analysis oftest results

Introduction

This chapter covers what is perhaps the most important subject in the book. Someof the original text carried over from the first edition may appear to some readersas being ‘old fashioned’. Furthermore many readers will not be acquainted with theBourdon pressure gauges mentioned in the text – these are devices in which the effectof the internal forces being measured can be seen to produce physical movement inthe instrument linkage.

It is the disconnection from close acquaintance with the forces and temperaturesbeing measured, induced by remote, ‘black box’ instrumentation, that is an importantjustification for this chapter. Modern instrumentation, data manipulation, simulationand ‘modelling’ have tended to obscure questions of accuracy and to give an illusionof precision to experimental results that is often totally unjustified. The intention ofthis chapter is to point out the dangers and call for a much greater awareness of theproblem.

It is a repeated statement of the obvious that the purpose of engine testing isto produce data and that the value of that data, particularly when transformed into‘information’, depends critically on its accuracy.

Appropriate accuracy in the test process presents the biggest challenge facing theexperimenter, who in addition to a complete understanding of the unit under test(UUT) must master the following skills:

• experience in the correct use of instruments;• knowledge of methods of calibration and an awareness of the different kinds of

error to which instruments are subject, both individually and collectively (as partof a complex system);

• a critical understanding of the relative merits and limitations of different methodsof measurement and their applicability to different experimental situations;

• an understanding of the differences between true and observed values of experi-mental quantities;

• an appreciation that repeatability and accuracy are separate and distinctly differentattributes of data;

The pursuit and definition of accuracy: statistical analysis of test results 409

• an appreciation of the effects on raw data of the signal processing chain;• an appreciation of data storage and retrieval, and relational database systems.

This is a very wide range of skills, only to be acquired by experience.The first essential is to acquire, as a habit of mind, a sceptical attitude to all

experimental observations: all instruments tend to be liars.

The basics

The digital display of a value gives an illusion of accuracy which can be totallyunjustified. The temptation to believe that a reading of, say, 97.12�C is to be reliedupon to the second decimal place is very strong, but in the absence of convincingproof must be resisted. An analogue indicator connected to a thermocouple cannotin most cases be read to closer than 1�C and the act of reading such an indicatoris likely to bring to mind the many sources of inaccuracy in the whole temperaturemeasuring system. Replace the analogue indicator by a digital instrument reading to0.1�C and it is easy to forget that all the sources of error are still present: the fact thatthe readout can now discriminate to 0.05�C does not mean that the overall accuracyof the measuring system is producing a reading within comparable limits.

Analogue display of data has a subtly different message to give to the observerin that it shows how far the measured point is from other values above and below itin addition to the rate and direction of change.

Example: measurement of exhaust temperature of a diesel engine

Temperature measurements are some of the most difficult to make accurately andthe following example illustrates many of the pitfalls encountered in the pursuit ofaccuracy. Figure 20.1 shows a typical situation: the determination of the exhaust tem-perature of a diesel engine. The instrument chosen is a vapour pressure thermometer,such as is suitable for temperatures of up to 600�C and commonly installed in theindividual exhaust ports of medium speed diesel engines. It comprises a steel bulb,immersed in the gas of which the temperature is to be measured, and connected bya long tube to a Bourdon gauge which senses the vapour pressure but is calibratedin temperature.

Let us consider the various errors to which this system is subject.Sensing errors are associated with the interface between the system on which the

measurements are to be made and the instruments responsible for those measure-ments. In the present case, there are a number of sources of sensing error. In the firstplace, the bulb of the temperature indicator can ‘see’ the walls of the exhaust pipe, andthese are inevitably at a lower temperature than that of the gas flowing in the pipe. Itfollows that the temperature of the bulb must be less than the temperature of the gas.This error can be reduced but not eliminated by shielding the bulb or by employinga ‘suction pyrometer’. A further source of error arises from heat conduction from the

410 Engine Testing

Figure 20.1 Measurement of exhaust gas temperature by vapour pressurethermometer

bulb to the support, as a result of which there is a continuous flow of heat from theexhaust gas to the bulb and no equality of temperature between them is possible.

A more intractable sensing error arises from the circumstances that the flow ofgas in the exhaust pipe is constant neither in pressure, velocity nor temperature.Pulses of gas, originating at the opening of the exhaust valves in individual cylinders,alternate with periods of slower flow, while the exhaust will also be to some extentdiluted by scavenge air carried over from the inlet. The thermometer bulb is thusrequired to average the temperature of a flow that is highly variable both in velocityand in temperature, and it is unlikely in the extreme that the actual reading willrepresent a true average.

A more subtle error arises from the nature of exhaust gas. Combustion will havetaken place, resulting in the creation of the exhaust gas from a mixture of air and fuel,perhaps only a few hundredths of a second before the attempt is made to measureits temperature. This combustion may still be incomplete, the effects of dissociationarising during the combustion process may not have worked themselves out, andit is even possible that the distribution of energy between the different modes ofvibration of the molecules of exhaust gas will not have reached its equilibrium value.As a consequence it may not be possible even in principle to define the exhausttemperature exactly.

We can deal with some of these sensing errors by replacing the steel bulbof Fig. 20.1 by a slender thermocouple surrounded by several concentric screens(Fig. 20.2) and further improve accuracy by the use of a suction pyrometer, Fig. 20.3,in which a sample of the exhaust gas is drawn past the sensor at uniform velocity byexternal suction. This deals with the ‘averaging’ problem but still leaves the questionregarding the definition of the exhaust temperature open.

Our example includes a Bourdon type pressure gauge and these may be particularlyprone to a variety of classic instrument errors.


Figure 20.2 Measurement of exhaust gas temperature by shielded thermo-couple

Figure 20.3 Measurement of exhaust gas temperature by suction pyrometer

Two of the most common are zero error and calibration error. The zero error ispresent if the pointer does not return precisely to the zero graduation when the gaugeis subjected to zero or atmospheric pressure.

Calibration errors are of two forms: a regular disproportion between the instrumentindication and the true value of the measured quantity, and errors that vary in anon-linear manner with the measured quantity. This kind of fault may be eliminatedor allowed for by calibrating the instrument, in the case of a pressure gauge by meansof a dead weight tester. These are examples of systematic errors.

In addition, the pressure gauge may suffer from random errors arising from frictionand backlash in the mechanism. These errors affect the repeatability of the readings.

The sensitivity of an instrument may be defined as the smallest change in appliedsignal that may be detected. In the case of a pressure gauge it is affected particularlyby friction and backlash in the mechanism. The precision of an instrument is definedin terms of the smallest difference in reading that may be observed. Typically,it is possible to estimate readings to within 1/10 of the space between graduations,

412 Engine Testing

provided the reading is steady, but if it is necessary to average a fluctuating readingthe precision may be much reduced.

Finally, one must consider the effect of installation errors. In the present casethese may arise if the bulb is not inserted with the correct depth of immersion inthe exhaust gas or, as is quite often the case, if it is installed in a pocket and isnot subjected to the full flow of the exhaust gas. If a thermocouple probe that isphysically smaller than the bulb is used, some of the vagaries of gas flow temperaturemeasurement may be greater.

A consideration of this catalogue of possible errors will make it clear that it isunlikely that the reading of the indicator will reflect the temperature of the gas in thepipe with any degree of exactness. It is possible to analyse the various sources of errorlikely to affect any given experimental measurement in this way, and while somemeasurements, for example of lengths and weight, require a less complex analysis,others, notably readings of inherently unsteady properties such as flow velocity,require to be treated with scepticism. A hallmark of the experienced experimenter isthat, as a matter of habit and training, he questions the accuracy and credibility ofevery experimental observation.

A similar critical analysis should be made of all instrumentation. This examplehas been dealt with in some detail to illustrate the large number of factors that mustbe taken into account.

Some general principles

1. Cumulative measurements are generally more accurate than rate measurements.Examples are the measurement of speed by counting revolutions over a periodof time and the measurement of fuel consumption by recording the time takento consume a given volume or mass.

2. Be wary of instantaneous readings. Few processes are perfectly steady. Thisis particularly true of flow phenomena, as will be clear if an attempt is madeto observe pressure or velocity head in a nominally steady gas flow. A watermanometer fluctuates continually, and at a range of frequencies depending onthe scale of the various irregularities in the flow. No two successive powercycles in a spark ignition engine are identical. If it is necessary to take a readinginstantaneously it is better to take a number in quick succession and statisticalanalyse them.

3. Be wary of signal damping, often entirely sensible; sometimes leading to seriousmisrepresentation of the true nature of a fluctuating value.

4. A related problem: ‘time slope effects’. If one is making a number of differentobservations over a period of nominally steady state operation, make sure thatthere is no drift in performance over this period.

5. Events must be recorded in the correct order for the correct causal deductions tobe made; the more complex the data streams and number of sources, the moredifficult this becomes.


6. The closer one can come to an ‘absolute’ method of measurement, e.g. pressureby water or mercury manometer or dead weight tester; force and mass by deadweights (plus knowledge of the local value of ‘g’ (see Chapter 8)), the less thelikelihood of error.

7. In instrumentation, as in engineering generally, simplicity is a virtue. Eachelaboration is a potential source of error.

Definition of terms relating to accuracy

This is a particularly ‘grey’ area.1–3 The statement ‘this instrument is accurate to±1 per cent’ is in fact entirely meaningless in the absence of further definition(Table 20.1).

Table 20.1 Various terms that are used, often incorrectly, in any discussion ofaccuracy

Error The difference between the value of a measurement asindicated by an instrument and the absolute or true value

Sensing error An error arising as a consequence of the failure of aninstrument to sense the true value of the quantity beingmeasured

Systematic error An error to which all the readings made by a given instrumentare subject; examples of systematic errors are zero errors,calibration errors and non-linearity

Random error Errors of an unpredictable kind. Random errors are due to suchcauses as electrical supply ‘spikes’ or friction and backlashin mechanisms

Observer error Errors due to the failure of the observer to read the instrumentcorrectly, or to record what he has observed correctly

Repeatability A measure of the scatter of successive readings of the samequantity

Sensitivity The smallest change in the quantity being measured that canbe detected by an instrument

Precision The smallest difference in instrument reading that it is possibleto observe

Average error Take a large number of readings of a particular quantity,average these readings to give a mean value and calculatethe difference between each reading and the mean. Theaverage of these differences is the average error; roughlyhalf the readings will differ from the mean by more than theaverage error and roughly half by less than the average error

414 Engine Testing

Statements regarding accuracy: a critical examination

Wemust distinguish between the accuracy of an observation and the claimed accuracyof the instrument that is used to make the observation. Our starting point must bethe true value of the quantity to be measured, Fig. 20.4a.

The following are a few examples of such quantities associated with enginetesting:

• fuel flow rate;• air flow rate;• output torque;• output speed;• pressures in the engine cylinder;• pressures in the inlet and exhaust tracts;• gas temperatures;• analysis of exhaust composition.

By their nature, none of these quantities is perfectly steady and decisions as to thesampling period are critical, see above.

Sensing errors

These are particularly difficult to evaluate and can really only be assessed by asystematic comparison of the results of different methods of measuring the samequantity.

Sensing errors can be very substantial. In the case discussed above regardingmeasurement of exhaust temperature, the authors have observed errors as great as70�C, Fig. 20.4b.

Typical sources of sensing error include

1. mismatch between temperature of a gas or liquid stream and the temperature ofthe sensor;

2. mismatch between true variation of pressure and that sensed by a transducer atthe end of a connecting passage;

3. inappropriate damping or filtration applied within the instrument’s measuringchain;

4. failure of a transducer to give a true average value of a fluctuating quantity;5. air and vapour present in fuel systems;6. time lag of sensors under transient conditions.

Note that the accuracy claimed by its manufacturer does not usually include anallowance for sensing errors. These are the responsibility of the user.


Mean

Time

Time

Time

Time

Time

Valu

eV

alu

eV

alu

eV

alu

eV

alu

e

True mean value

Mean valueindicatedby instrument

True value

Mean valueindicatedby instrument

Value sensedby instrument

Value indicatedby instrument

Value sensedby instrument

(a) True value of quantity to be measured

(b) Sensing error

(c) Systematic error

(d) Random error

(e) Instrument drift

Figure 20.4 Various types of instrumentation error. (a) True value of quantityto be measured plotted against time – on the same time scale; (b) sensing error;(c) systematic error; (d) random error; (e) instrument drift

416 Engine Testing

Systematic instrument errors (Fig. 20.4c)

Typical systematic errors include the following:

1. Zero errors: the instrument does not read zero when the value of the quantityobserved is zero.

2. Scaling errors: the instrument reads systematically high or low.3. Non-linearity: the relation between the true value of the quantity and the indicated

value is not exactly in proportion; if the proportion of error is plotted againsteach measurement over full scale, the graph is non-linear.

4. Dimensional errors: for example, the length of a dynamometer torque arm maynot be precisely correct.

Random instrument errors

Random errors, Fig. 20.4d include

1. effects of stiction in mechanical linkages;2. effects of friction, for example in dynamometer trunnion bearings;3. effects of vibration or its absence;4. effects of electromagnetic interference (see Chapter 10).

Instrument drift

Instrument drift, Fig. 20.4e, is a slow change in the calibration of the instrument asthe result of

1. changes in instrument temperature or a difference in temperature across it;2. effects of vibration and fatigue;3. fouling of the sensor, blocking of passages, etc.;4. inherent long-term lack of stability.

It will be clear that it is not easy to give a realistic figure for the accuracy of anobservation.

Instrumental accuracy: manufacturers’ claims

With these considerations in mind it is possible to look more critically at the statement‘the instrument is accurate to within ±1 per cent of full scale reading’.

At least two different interpretations are possible:

1. No reading will differ from the true value by more than 1 per cent of full scale.This implies that the sum of the systematic errors of the instrument, plus thelargest random error to be expected, will not exceed this limit.


2. The average error is not more than 1 per cent of full scale. This implies thatabout half of the readings of the instrument will differ from the average byless than 1 per cent full scale and half by more than this amount. However, thedefinition implies that systematic errors are negligible: we are only looking atrandom errors.

Neither of these definitions is satisfactory. In fact the question can only be dealt withsatisfactorily on the basis of the mathematical theory of errors.

Uncertainty

While it is seldom mentioned in statements of the accuracy of a particular instrumentor measurement, the concept of uncertainty is central to any meaningful discussionof accuracy.

Uncertainty is a property of a measurement, not of an instrument:

The uncertainty of a measurement is defined as the range within which the true value islikely to lie, at a stated level of probability.

It is interesting to note that in most scientific disciplines data, when presentedgraphically, have each data point represented by a cruciform indicating the degreeof uncertainty or error band; this practice is rarely, if ever, seen in engine test dataand helps to confirm the unjustified illusion of exactness.

The level of probability, also known as the confidence level, most often used inindustry is 95 per cent. If the confidence level is 95 per cent, there is a 19 to 1 chancethat a single measurement differs from the true value by less than the uncertaintyand one chance in 20 that it lies outside these limits.

If we make a very large number of measurements of the same quantity and plotthe number of measurements lying within successive intervals, we shall probablyobtain a distribution of the form sketched in Fig. 20.5. The corresponding theoretical

Value

Num

ber

Figure 20.5 A frequency distribution

418 Engine Testing

Mean

Standard deviation

Average error ormean deviation

Value

95% confidencelevel

1.96σ

σ

Figure 20.6 Normal or Gaussian distribution

curve is known as a normal or Gaussian distribution, Fig. 20.6. This curve is derivedfrom first principles on the assumption that the value of any ‘event’ or measurementis the result of a large number of independent causes (random sources of error).

The normal distribution has a number of properties shown on Fig. 20.6:

• The mean value is simply the average value of all the measurements.• The deviation of any given measurement is the difference between that measure-

ment and the mean value.• The �2 (sigma)2 is equal to the sum of the squares of all the individual deviations,

divided by the number of observations.• The standard deviation � is the square root of the variance.

The standard deviation characterizes the degree of ‘scatter’ in the measurements andhas a number of important properties. In particular, the 95 per cent confidence levelcorresponds to a value sigma = 1.96. Ninety-five per cent of the measurements willlie within these limits and the remaining 5 per cent in the ‘tails’ at each end of thedistribution.

In many cases the ‘accuracy’ of an instrument as quoted merely describes theaverage value of the deviation, i.e. if a large number of measurements are madeabout half will differ from the true or mean value by more than this amount andabout half by less. Mean deviation = 0.8� approximately.

However, this treatment only deals with the random errors: the systematic errorsstill remain. To give a simple example, consider the usual procedure for checking thecalibration of a dynamometer torque transducer. A calibration arm, length 1.00m,carries a knife-edge assembly to which a dead weight of 10 kg is applied. The loadis applied and removed 20 times and the amplifier output recorded. This is found torange from 4.935 to 4.982 volts with a mean value 4.9602 volts.


At first sight we could feel a considerable degree of confidence in this mean valueand derive a calibration constant:

k=4�9602

10×9�81×1�0= 0�050563V/Nm�

The 95 per cent confidence limit for a single torque reading may be derived from the20 amplifier output readings and, for the limiting values assumed, would probablybe about ±0.024V, or ±0.48 per cent, an acceptable value.

There are, however, four possible sources of systematic error:

1. The local value of g may not be exactly 9.81m/s2 (see Chapter 8).2. The mass of the dead weight may not be exactly 10 kg.3. The length of the calibration arm may not be exactly 1.00m.4. The voltmeter used may have its own error.

In fact none of these conditions can ever be fulfilled with absolute exactness. Wemust widen our 95 per cent confidence band to take into account these probablyunknown errors.

This leads straight on to our next topic.

Traceability

In a properly organized test department, careful thought must be given to this matter.In many small departments the attention given to periodic recalibration is minimal.In many more, while some effort may be made to check calibrations from time totime, the fact that the internal standards used for these checks may have their ownerrors is overlooked. To be confident in the validity of one’s own calibrations, thesestandards must themselves be checked from time to time.

Traceability refers to this process. The ‘traceability ladder’ for the dead weightreferred to in the above example might look like this:

• the international kilogram (Paris);• the British copy (NPL);• national secondary standards (NPL);• local standard weight (portable);• our own standard weight.

Combination of errors

Most derived quantities of interest to the experimenter are the product of severalmeasurements, each of them subject to error. Consider, for example, the measurement

420 Engine Testing

of specific fuel consumption of an engine. The various factors involved and typical95 per cent confidence limits are as follows:

• torque ±0.5 per cent;• number of revolutions to consume measured mass of fuel ±0.25 per cent;• actual mass of fuel ±0.3 per cent.

The theory of errors indicates that in such a case, in which each factor is involvedto the power 1, the confidence limit of the result is equal to the square root of thesum of the squares of the various factors, i.e. in the present case, the 95 per centconfidence limit for the calculated specific fuel consumption is

√

0�52+0�252+0�32 =±0�58%

The number of significant figures to be quoted

One of the most common errors in reporting the results of experimental work isto quote a number of significant figures, some of which are totally meaningless –the illusion of accuracy once more; often engendered by the misuse of a commonspreadsheet format. This temptation has been vastly increased by the now universalemployment of digital readouts.

Consider a measurement of specific fuel consumption readings as follows:

engine torque 110.3Nmnumber of revolutions 6753mass of fuel 0.25 kg

Specific consumption equals

3�6×106×0�25

2×�×6753×110�3= 0�1923kg/kWh

Let us assume that each of the three observations has the 95 per cent confi-dence limit specified in the previous paragraph. Then the specific fuel consumptionhas a confidence limit of 0.58 per cent, i.e. between 0.1912 and 0.1934 kg / kWh.Clearly it is meaningless to quote the specific fuel consumption to closer than0.192 kg / kWh.

Particularly in development work where the aim is often to detect smallimprovements this question must be kept in mind.


Absolute and relative accuracy

This is a topic linked with the last one. In a great deal of engine test work – perhapsthe majority – we are interested in measuring relative changes and the effect ofmodifications. The absolute values of a parameter before and after the modificationmay be of less importance; it may not be necessary to concern ourselves with theabsolute accuracy and traceability of the instrumentation. Sensitivity and precisionmay be of much greater importance. Some test engineers have an unhealthy respectfor the attribute of test cell ‘repeatability’ of results while not considering that thiscan be almost completely disconnected from accuracy of results.

The cost of accuracy: a final consideration

An engineer with responsibility for the choice of instrumentation should be awareof the danger of mismatch between what he thinks he requires and what is actuallynecessary for an adequate job to be done. The cost of an instrument to perform aparticular task can vary by an order of magnitude; in most cases the main variableis the level of accuracy offered and this should be compared with the accuracy thatis really necessary for the job in hand.

Summary

The degree of understanding of the subject of accuracy is perhaps the main criterionby which the professional quality of a test engineer should be judged. A properunderstanding requires a wide range of knowledge and the principal elements of thishave been described, with a special warning as to the dangers of digital readouts. Anexample of temperature measurement illustrates the problems and terms relating toaccuracy have been defined.

This leads to a discussion of the meaning of the term as used by instrumentmanufacturers, and different types of error are identified. The mathematical basis ofthe concept of uncertainty is given and the method of dealing with combined errorsis described.

References

1. Hayward, A.T.J. (1977) Repeatability and Accuracy, Mechanical Engineering Publications,London. ISBN: 0-7008-0491-9.

2. Dietrich, C.F. (1973) Uncertainty, Calibration and Probability, Adam Hilger, London.3. Campion, P.J. et al. (1973) A Code of Practice for the Detailed Statement of Accuracy,

HMSO, London.

422 Engine Testing

Further reading

BS 4889 Method for Specifying the Performance of Electrical and Electronic Mea-

suring Equipment.

BS 5233 Glossary of Terms used in Metrology.

BS 5497 Part 1 Guide for the Determination of Repeatability and Reproducibility

for a Standard Test Method by Inter-laboratory Tests.

BS 7118 Part 1 and 2 Linear and Non-linear Calibration Relationships.

http://www.npl.co.uk/force/faqs/glossary.htmlhttp://www.swan.ac.uk/lis/help_and_training/pdf/standards.pdf

Index

4 x 4 dynamometers, 384–516 cylinder V engines, 176, 17730-30-30-10 rule, 16, 1750 to 450 kW automotive engines, 50–3100 kW engines, 16, 75–7100 kW gasoline engines, 16250 kW diesel engines, 273–5250 kW engines, 90–4, 273–5250 kW test cells, 121250 kW turbocharged engines, 273–5

Above-floor dilution ducts, 126Above ground fuel tanks, 130–1ABS brake tests, 372–3Absolute accuracy, 421Absolute pressure relationships, 234Absorption coefficients, 40Acceleration measurement, 152–3, 232, 236Acceptance tests, 315Accuracy:

absolute, 421analogue-to-digital conversion, 399basics, 409–12chassis dynamometers, 391–3cost, 421definitions, 408, 413error terms, 413instrumentation, 412–13, 415–19relative, 421results, 408–22significant figures, 420statements, 414–16see also Errors

a.c. dynamometers, 158–9, 207–8Acoustic tiles, 40ADC see Analogue-to-digital conversionAdiabatic engines, 263Administrative tasks, 308–23

Aftermarket emission legislation, 335After shut-down checks, 221After start-up checks, 221AHU see Air handling unitsAir:

properties, 250–1specific enthalpy, 96

Airbox method, 256–60Air brake dynamometers, 160Air charge mass, 253–5Air conditioning, 72–107

combustion air, 99–105cooling, 97–8humidity warning, 98–9Legionnaires’ disease, 99notation, 106–7processes, 96psychometry, 94–9purpose, 72

Air consumption, 251airbox method, 256–60calculations, 259measurement, 242–62

Air flow, 91–2, 256–7, 275Air/fuel ratio, 286–7, 327Air handling units (AHU), 73–4, 85Air mass flowmeters, 260Air-springs, 33Air standard cycle efficiency, 266, 267–8Air temperature, 105, 253–6Air treatment see Air conditioningAlarms:

computer monitoring, 218–19interaction matrix, 212nominated tables, 229

Alternating current see A.c.Ambient conditions guidelines, 199American...: see also United States...

424 Index

American Society for Testing and Materials(ASTM) standards, 354, 356–9,366, 367

Amplitude relationships, 27, 172, 173Analogue-to-digital conversion (ADC)

accuracy, 399Analysis:

gaseous emissions, 337–43software, 17, 18test programmes, 313–14test sequences, 270–1see also Combustion analysis

Anechoic tests, 21, 34, 40–2, 63–4,376–7, 404

Anemometers, 261Annular type rubber couplings, 189Antifreeze, 109, 115Arrangements: test cells, 50–3‘As built’ electrical documentation/drawing

standards, 212ASTM see American Society for Testing

and Materials standardsAsynchronous dynamometers, 158–9ATEX (Atmospheric Explosion)

codes, 67–8Atmospheric pressure, 251–2Authority: European tests, 365Automotive engines, 50–3, 55, 368–94AVL dynamic combustion air pressure

control, 105Axial flow fans, 89Axial shock loading, 190

Balance see Energy balanceBalanced fan systems, 84, 85Balance sheets, 14, 15, 16Bearings: couplings, 171Bedplates, 30–4Before start-up checks, 221Benchmarking, 4Bench tests: tribology, 356–7Biofuels, 363Black body emission, 76Blades: fans, 89Blow-by meters, 250BMEP see Brake mean effective pressureBMS see Building management systemsBolt-on dynamometers, 49Bolt-on variable fill machines, 157

Boom boxes, 51–2Bore polishing, 249, 356, 365Bourdon gauges, 408, 410–11Brake mean effective pressure (BMEP),

277, 280Brake testers, 372–3Breeze block absorption coefficients, 40Brushless torque-shafts, 148BSI standards, 316–17, 366, 367Building management systems (BMS),

219–20Buildings: exhaust gas cowls, 125–7Bulk fuel storage, 129–31Burn rate curves, 288, 290, 306

Cabinet ventilation, 210Cables, 201–10

capacitive interference, 205chassis dynamometers, 386electromagnetic interference, 206–7fire stopping, 209–10inductive interference, 204–5layouts, 203–7shielded, 201–2spacing, 206–7trays/trunking, 205

Calibration:chassis dynamometers, 391–3computerized procedures, 238dynamometers, 150–1engines, 395, 400–1gaseous emissions analysers, 340–2high end engines, 301–2signal chains, 229–30torque measurement, 149–52

Calorific value: fuels, 264Calorimeters, 271–2, 281, 288, 289CAN (Control Area Network) bus, 302,

383, 402Capacitive interference, 205Carbon, 325Carbon dioxide, 69Carbon monoxide, 324, 325Cardan shafts, 181CDM regulations, 9, 312CE (Conformité Europeen) marking, 211CEC publications/tests, 364–5Cells see Test cellsCentrifugal fans, 87, 89

Index 425

Certification:emissions, 329–31, 341lubricants, 354–5

CFO see Critical flow orificeCFR see Co-operative fuel researchChange notification, 9–10Characteristics see PropertiesCharge see Turbocharged enginesCharge amplifier circuits, 296–7Charts:

psychometric, 95recorders, 398

Chassis dynamometers, 368–94accuracy, 391–3cables, 386calibration, 391–3drive cabinet housing, 385emission testing, 374–5fire suppression, 389four roller type, 375, 380–1guards, 387–8installation, 381limitations, 393–4offloading machines, 385pit design, 381–4positioning machines, 385road load equation, 369–71rolls, 388, 390–1safety, 387–8single roll set, 380specification, 391–3wheel substitution type, 381wind tunnels, 378see also Rolling road dynamometers

Checks:after shut-down, 221after start-up, 221before start-up, 220

Chemiluminescence detector (CLD), 337–8Chemistry: engine emissions, 325–6Choice:

couplings, 181–2dynamometers, 154, 166–9emissions testing sites, 350–1instrumentation, 231software, 227, 406–7transducers, 231

Classifications:dynamometers, 154–62emissions legislation, 332–6

fans, 89–90lubricants, 354–5

CLD see Chemiluminescence detectorClimatic test cells, 377–80Clutches, 191CNG see Compressed natural gasCold testing, 55–6, 403–4Columns: cooling water, 115Combination of errors, 419–20Combined spring/rubber mountings, 32Combustion, 282–307

air consumption, 242–62air/fuel ratio, 286–7constant volume calorimeter, 288, 289crank angle diagrams, 288, 289cylinder pressure diagrams, 291, 294definitions, 285diesel engines, 284–93, 318energy release, 292flame ionization detectors, 287fundamentals, 283gaseous components, 325gasoline engines, 283–4‘knock’ sensing, 284, 303mass fraction burned, 288, 289, 290–2power output variation, 286–7pressure measurement, 288, 289, 296–9processes, 325raw measured data, 305–6specific fuel consumption, 286–7top dead centre determination, 300

Combustion air:AVL pressure control systems, 105centralized supply, 100condensate drain lines, 104conditioning, 251engine performance, 251pressure control, 104–5supplies, 100system contents, 102treatment, 99–105units, 100–1, 104

Combustion analysis, 282–307charge amplifier circuits, 296–7data storage, 282occasional analysis, 302–3result calculation, 304–6workflow process, 297–8

Commissioning cooling water circuits,119–20

Common or individual services?, 19–20

426 Index

Communication, 11, 402Compounds: in water, 112Compressed natural gas (CNG), 134, 264

see also Natural gasCompression ratio effect, 265–8Compression rings, 360–1Computers:

alarm signal monitoring, 218–19calibration procedures, 238data acquisition, 399–400data recording, 398–9engine indicating, 296test cell role, 228–33

Concrete:absorption coefficients, 40pits, 33planks, 60

Condensate drain lines, 104Conductive coupling interference, 207Confidence levels, 417, 418Confidential information, 4Conformité Europeen (CE) marking, 211Conformity, 67, 336Connecting rod crank mechanism, 22Consolidated subsoils, 33Constant fill machines, 155Constant volume calorimeters, 288, 289Constant volume sampling (CVS) systems,

343–6, 350–1Constraints: projects, 5, 9, 11Consumption measurement:

combustion air, 242, 250, 256–60fuel, 242–62gaseous fuels, 247notation, 261oil, 242, 248–9rate meters, 245–7

Contractors, 8–9Contract record sheets, 9–10Control:

dynamic engine testing, 239–41endurance testing, 238heat exchanger strategies, 118humidity, 101–4oil temperature units, 140software choice, 227system specifications, 6test cells, 15, 64–6, 216–41, 402test sequences, 221–8thermal shock, 238–9

transient testing, 239–41unmanned running, 238ventilation systems, 87

Control Area Network (CAN) bus, 302,383, 402

Control desks/consoles, 50–3, 61, 63, 65–6,69, 217

Controlled shutdown/stop, 217Control modes:

four-quadrant dynamometers, 225nomenclature, 221–2position, 222–3problems, 226–7test sequences, 221–7

Control rooms, 47–71, 92–4Control surfaces, 15Control volume:

250 kW turbocharged dieselengines, 274

internal combustion engines, 268–9test cells, 14–15

Conversion factors xv-xviCoolants, 115–18

see also Air; OilCooling: moisture content reduction, 97–8Cooling air heat capacity, 74–5Cooling columns, 115Cooling systems:

columns, 115coolants, 115–18engine oil, 140exhaust gases, 125

Cooling towers, 19, 99, 111, 113–14Cooling water, 108–28

circuit types, 113–20closed plant circuits, 114–15columns, 115commissioning circuits, 119–20flow, 275flow rates, 109–10flow velocities, 118heat exchangers, 116–19loads, 110open plant circuits, 113–14‘raw’ circuits, 113–15sumps, 113–14supply systems, 108

Co-operative fuel research (CFR), 141Coriolis effect flowmeters, 247Correction factors, 315, 316–18

see also Accuracy; Errors

Index 427

Cost of accuracy, 421Couplings, 170–96

annular type, 189background reading, 171–2bearings, 171cardan shafts, 181choice, 181–2damping, 182–5design summary, 193drive line, 171, 191–2drive shaft design, 185–93dynamic magnifier, 185elastomeric element type, 182features, 170–1flexible, 182–5guards, 193keyways, 179–80multiple disc flexible type, 182notations, 195–6quill shafts, 181rubber bush type, 182–3, 189service factors, 186, 188shafts, 178–9, 193shock loading, 189–90stress concentrations, 179–80torque capacity, 190torsional stiffness, 191

Cowls on buildings, 125–7Cradle machines, 144–5, 146Crank angles, 288, 289, 292, 304Crankcase blow-by measurement, 249–50Cranking engines, 164–6, 189–90Critical flow orifice (CFO), 341, 344Critical speeds: torsional oscillations, 172–8Cumulative flowmeters, 243–5Customer relations, 7–11Cutter FID, 338–9CVS systems see Constant volume

sampling systemsCycles see Standard cyclesCyclic energy release, 293–5Cyclic force, 233Cylinder pressure diagrams, 291, 294

Dampers: fire, 86Damping:

couplings, 182–5ratios, 28

Dangerous substances regulations, 129

Data:chart recorder format, 398computer recording, 398–9flow, 397post-test processing, 395security, 406

Data acquisition see Data collectionData collection, 216–41, 395

computers, 399–400engine mapping, 400–1system specifications, 6traditional approach, 395–7transducer chains, 229

Data handling, 395–407Data management, 404–5Data processing, 403–4, 405Data recording, 398–9Data storage, 282Data transfer protocols, 401–2Day-tank systems, 135d.c. dynamometers, 158Decelerating conditions, 152–3Degrees of freedom: internal combustion

engines, 22Department organization, 308–23Depth of pits, 382–3Design:

control rooms, 47–71couplings, 193drive shafts, 185–93ducts, 80–3dynamometers, 384–5earthing systems, 202–3engine mounting, 24–30experiments, 308, 318–22pits, 381–4test cells, 47–71ventilation systems, 90–4see also Electrical design

Designated engines, 359–62Detonation: ‘Knocking’, 284, 303Deviation: measurement, 314, 317, 418Dew point, 23, 94, 270Diesel engines:

250kW turbocharged, 90–4combustion systems, 284–93, 318emissions, 328–9, 333–4energy balance, 272, 273–5exhaust temperature measurement,

409–12experiment design, 318–22

428 Index

Diesel engines (cont.)experiment results, 320–1fuel properties, 142heat release curves, 291, 294large-/medium-speed, 56–7legislation, 333–4single cylinder, 256–7, 291, 294thermal efficiency, 268time scales, 293Willan’s line, 279see also Engines

Diesel fuels, 367Differential pressure relationships, 234Dilution tubes, 127Dimensions: test cells, 48Direct current dynamometers, 158Direct injection diesel engines, 284,

318, 328Directives: EEC, 209Direct weighing fuel gauges, 244Disaster recovery, 405Disc type dynamometers, 147, 157–8Displacement measurement, 235–6Display parameters, 302Distributed input/output, 230Distribution systems: storing gas, 340–2Disturbing device interconnection, 208–9Diversity factors: energy balance, 17–19Documentation, 13, 212Doors: test cells, 58–9Drag: vehicles, 371Drawing standards, 212Drift, 298, 416Drive cabinet housing, 385Drive cycles, 331, 347–8Drive line: couplings, 171, 191–2Driver aids, 389Drive shaft design, 185–93Drive train calibration, 239Driving schedules, 349Drum stores: fuel, 131Dry-bulb temperatures, 96Dry gap dynamometers, 147Dry powder extinguishing systems, 70Dry sump operation, 248–9Ducting, 80–3, 86Dynamic amplifiers, 172, 173Dynamic engine testing control, 239–41Dynamic magnifiers, 185Dynamic test beds, 54–5

Dynamic viscosity, 357Dynamometers:

a.c. systems, 207–8advantages, 167–8air brake type, 160anechoic chamber type, 376–7bolt-on type, 49calibration, 150–1choice, 154, 166–9classification, 154–62d.c. systems, 158disadvantages, 167–8disc type, 147, 157–8eddy current type, 147, 159–60, 163–4electrical motor-based, 158–60engines

alignment, 192couplings, 170–96torque curves, 162–3

exhaust emission test procedure, 349friction type, 160Froude type, 145, 156heat sources, 79hybrid type, 160–1hydraulic, 154–8, 163hydrostatic, 158independent wheel type, 380–1matching engine characteristics, 162–4mileage accumulation, 375–6non-electrical starting systems, 166operating quadrants, 153–4, 161–2performance curves, 163–4Schenck type, 145, 147sluice-gate type, 145stands, 51, 52tandem type, 160–1torque measurement, 144–69trunnion-mounted machines, 144–5types, 167–8variable geometry type, 384–5see also Chassis dynamometers; Rolling

road dynamometers

Earthing systems design, 202–3Earth loops, 202ECE 15 cycle, 347–8Eddy current dynamometers, 147, 159–60,

163–4Edging: pits, 383–4

Index 429

Editing: test sequences, 228EEC:

CE marking, 211directives, 209see also European Union

Efficiency see Thermal efficiencyEGR see Exhaust gas recirculationEIPTs see Engine indicating pressure

transducersElastic curves, 176, 177Elastomeric element couplings, 182Electrical cabinet ventilation, 210Electrical design, 197–215

electrical installation, 198engineer’s role, 197–8notation, 213–15test cell environment, 199

Electrical documentation/drawingstandards, 212

Electrical installation, 198Electrical interference, 199–202, 204–5,

206–7Electrical motor-based dynamometers,

158–60Electrical power see Power supplyElectrical signals:

ingress protection rating, 200measurement interference, 199–202

Electromagnetic compatibility (EMC),202–3, 376–7

Electromagnetic interference (EMI),199–202, 206–7

EMC see Electromagnetic compatibilityEmergency brakes, 386Emergency stop, 66, 216–17EMI see Electromagnetic interferenceEmissions, 324–53

chassis dynamometers, 374–5diesel engines, 328–9emission benches, 342–3evaporative, 349–50internal combustion engines, 325–6legislation, certification and test

processes, 329–31particulate measurement, 336spark ignition engines, 326–8testing, 329–31, 374–5see also Exhaust emissions

Emissivity, 76, 107End-of-line production testing, 373–4

Endurance testing, 238Energy balance:

30-30-30-10 rule, 16, 17engines, 16–17internal combustion engines, 268–75kW per kW power output, 273predictions, 273–6rules of thumb, 16test cells, 108, 120–1turbocharged engines, 275typical values, 272–3

Energy equations: steady flow, 268, 269Energy flows:

250 kW turbocharged dieselengines, 274

100 kW gasoline engines, 16diagrams, 121hydraulic dynamometers, 16internal combustion engines, 268–9test cells, 16, 121

Energy release, 292, 293–6Engine indicating (EI), 282

computers, 296display parameters, 302equipment integration, 301hardware suppliers, 282software suppliers, 282see also Combustion; Combustion

analysisEngine indicating pressure transducers

(EIPTs), 303–4Engine-mounted starting systems,

165–6Engines:

16 cylinder V engines, 176, 177100 kW, 16, 75–7250 kW turbocharged diesel, 90–4airbox connection, 260air condition, 251air consumption, 251calibration, 395, 400–1clutches, 191complex test sequences, 228coolants, 115–18cranking, 164–6designated engines, 359–62dynamometer alignment, 192dynamometer couplings, 170–96emissions chemistry, 325–6energy balance, 16–17, 268–75first/second order forces, 23

430 Index

Engines (cont.)friction losses, 360fuel

condition, 247–8consumption, 322pressure controls, 137–8supply regulation, 138temperature controls, 138–40

governed, 224handling systems, 61–2hearing/seeing, 57–8as heat sources, 78heat transfer, 75–7‘knock’, 284, 303large, 177maps, 322, 395, 400–1matching dynamometers to, 162–4mechanical loss measurements, 276–80models, 395, 401, 404mounting, 24–30multi-cylinder, 23no starter motor, 164–5oil cooling systems, 140oil temperature control, 115–18performance, 247–8, 251portable stands, 49power codes, 317–18running away control, 66, 222, 223shut down control, 66speed map, 322stands, 49starting, 164–6stoichiometric, 255, 326temperature controls, 138–40test software, 406–7thermal shock testing, 119torque, 162–3, 322tribology test examples, 360–1see also Automotive...; Diesel...; Internal

combustion...; Spark ignition...; Testcells

Environmental legislation, 324, 329Environmental Protection Agency (EPA)

US standards, 325, 330, 334, 349, 350,373, 374

Equipment integration, 301Errors:

assessment, 149–52combination, 419–20instrumentation, 415–19

sensing, 414terms, 413torque measurement, 149–52see also Accuracy

Ethanol, 363Ethylene glycol (antifreeze), 109, 115EU see European UnionEUDC see Extra urban driving cycleEurope: test authority, 365European ATEX codes, 67–8European standards: safety, 211European Union (EU):

CVS systems, 344, 345emissions legislation, 329, 332–3exhaust emissions test procedure, 347–8SHED legislation, 350see also EEC

Evaporative emissions, 349–50Excess air factor (lambda ratio), 284, 285Exhaust calorimeters, 271–2, 281Exhaust emissions, 324–53

air/fuel ratio relation, 327constant volume sampling systems,

343–6diesel engines, 333–4instrumentation integration, 340legislation, 329–31, 332–6marine engines, 334–5motorcycles, 349notation, 352test procedure, 329–31, 347–9United States, 348–50see also Emissions

Exhaust gases:above-floor dilution ducts, 126cooling, 125cowls on buildings, 125–7diesel engines, 275, 409–12dual use, 125extraction ducts, 125flow, 275heat transfer, 77individual close coupled cells, 123–4loss measurements, 270multiple cells, 124–5purge systems, 125, 126subfloor collection, 126systems, 122–7temperature measurement, 409–12test cells, 122–7

Exhaust gas recirculation (EGR), 328

Index 431

Exhaust noise, 42–3Exhaust resonators, 42, 43, 44Exhaust systems see Exhaust gasesExperiment design, 308, 318–22Experimenter skills, 408–9Explosion risk, 67–8, 73Explosive atmospheres, 129External ducting, 86External noise, 39Extinguishing systems, 68, 70, 71Extraction ducts, 125Extra urban driving cycle

(EUDC), 348

Facility specifications, 1, 2–3see also Test cells

Fans, 87–94advantages, 90balanced, 84, 85blades, 89centrifugal, 87, 89classification, 89–90disadvantages, 90ducting, 83noise, 89plenums, 84spot fans, 86supplementary cooling, 86types, 90ventilation systems, 83

Fast FID, 338–9Feasibility studies, 3–4FIDs see Flame ionization detectorsFiltering: electrical, 299Final specifications, 17–19Fire:

cable penetrations, 209–10chassis dynamometers, 389dampers, 86suppression systems, 69

Fire control, 66–71carbon dioxide, 69extinguishing systems, 68inert gas, 69suppression systems, 69

Fittings: duct design, 81–3Flame ionization detectors (FIDs), 287,

338–9Fletcher–Munson curves, 37

Flexible couplings, 182–5Flexible mountings, 25Flooded inlets, 101Flooring, 58, 383–4Flow:

air flow, 91–2, 256–7, 275cooling water rates, 109–10cooling water velocities, 118data, 397diesel engine fluids, 274–5sharp-edged orifice, 256–7, 259see also Energy flows

Flowmeters:air mass, 260coriolis type, 247cumulative type, 243–5mass flow, 245–7positive displacement, 261

Flywheels, 193–5FMEP see Friction mean effective pressureFoam extinguishing systems, 70Force measurement: cyclic/quasistatic, 233Format: chart recorders, 398Forms: risk assessment, 311Foundations:

consolidated subsoils, 33isolated blocks, 33massive, 30–4subsoils, 33test beds, 24–9

Fourier transform infrared analyser(FTIR), 337

Four-quadrant dynamometers, 225Four-stroke engines, 270, 276–7

see also Internal combustion enginesFrequency distribution, 417Frequency relationships, 27, 172, 173Friction dynamometers, 160Friction losses, 360Friction mean effective pressure (FMEP),

278, 280Front end octane number (R100), 141Froude dynamometers, 145, 156FTIR see Fourier transform infrared

analyserFTP-75, 348–9Fuels, 129–43

air/fuel ratio, 286–7, 327calorific value, 264condition, 247–8consumption, 139, 242–62, 322

432 Index

Fuels (cont.)day-tank system, 135diesel engines, 142, 274drum stores, 131engine maps, 322engine performance, 247–8flows, 274gaseous, 247, 264–8in-cell systems, 136–7injection systems, 245, 327, 328–9lines, 133–4mixture preparation, 142octane numbers, 141pipes, 132pressure controls, 137–8properties, 142reference drums, 134reference fuels, 358storage, 129–43sulphur content, 328supply systems, 129–31, 134–6tank systems, 130–1, 135temperature control, 138–40testing, 354, 359–62underground lines, 133–4see also Gaseous fuels; Liquid fuels

Full anechoic test cells, 41Full power energy balances, 272, 273–5Functional specifications, 2, 7

Gantt charts, 12Gas:

analysers, 342–3flow measurement, 250

Gaseous components: combustion, 325Gaseous emission analysis/measurement,

337–43see also Emissions; Exhaust gases

Gaseous fuels, 247, 264–8Gasoline:

properties, 141, 366shelf life, 141standards, 366test methods, 366

Gasoline engines:combustion, 283–4emissions legislation, 333energy balance, 270, 272exhaust emissions, 327

Hook curves, 285, 286see also Internal combustion engines

Gas systems see Exhaust gasesGauges, 243–5Gaussian distribution, 418GC-FID, 338–9Glass: absorption coefficients, 40Gleichzeitigkeits Faktor, 17GMEP see Gross mean effective pressureGoodman diagrams, 179–80Governed engines, 224, 317Gravimetric fuel consumption gauges, 243,

244–5‘Green’ engines, 165, 166, 190Grids: earthing, 202–3Gross mean effective pressure (GMEP),

293–5Ground(ing)/earth(ing), 202–3Ground loops, 299–300Group roles, 308–9Guards, 193, 387–8

Hall effect transducers, 232, 236Halon extinguishing systems, 70Handling engines, 61–2Hardness: water, 111–12‘Hardware in loop’ testing, 401Hardware suppliers, 282Harmonic components, 173, 174Harmonic distortion, 199, 201Harshness see Noise, vibration and

harshnessHeader tanks, 116–17Health and safety, 129, 308, 309–12, 346Hearing engines in test cells, 57–8Heat capacity: cooling air, 74–5Heat exchangers, 116–19Heat losses, 79, 263–81Heat release curves, 291, 294Heat sources: test cells, 78–80Heat transfer:

engines, 75–7exhaust systems, 77ventilation air, 79walls, 77–8

Height variation: atmospheric pressure, 252Hemi-anechoic test cells, 41High end engine calibration, 301–2High level inlet ducting, 84

Index 433

High-speed see Large-speedHomologation, 335–6Hook curves, 285, 286Hotfilm/wire anemometers, 261Hot testing, 55, 56Hub connections, 179–80Humidity:

air charge mass, 253–5control units, 101–4load calculations, 98operational envelope specification, 100warning, 98–9

Hybrid dynamometers, 160–1Hydraulic dynamometers, 154–8, 163

elastic curves, 176, 177energy flows, 16

Hydrocarbons, 324, 326Hydrokinetic dynamometers, 154–8

‘bolt-on’ machines, 157constant fill machines, 155disc type, 157–8variable fill machines, 156–7

Hydrostatic dynamometers, 158

IBGT see Insulated bipolar gate transistorsIC engines see Internal combustion enginesIdeal standard cycles, 265–8IMEP see Indicated mean effective pressureIn-cell controls, 65–6In-cell fuel systems, 136–7Inclined test beds, 54–5Independent wheel dynamometers, 380–1Indicated mean effective pressure (IMEP),

277, 280, 293–5Indicated power, 273, 277–8Indication equipment, 299–300Indirect injection diesel engines, 284, 328Individual close coupled cells, 123–4Induction air flow, 275Inductive interference, 204–5Inductive transducers, 235–6Inergen extinguishing systems, 70Inert gas fire control, 69Inertia: flywheels, 195Inflows: test cells, 15Ingress protection rating, 200Injection systems: fuel, 245, 327, 328–9Inlets, 83–5, 101In-line shaft torque measurement, 145–8

In-service assessment/tuning, 373Inspection and maintenance, 61, 123, 145,

220–1, 358, 359Installation: dynamometers, 381, 384–5Instantaneous energy release, 292, 294,

295–6Institute of Petroleum (IP) standards,

366, 367Instrumentation:

accuracy, 412–13, 415–19choice, 231drift, 416errors, 412–13, 415–19exhaust emissions, 340integration, 340liquid fuel consumption, 242–7measurement, 232, 238, 242–7operational, 217primary, 65safety, 217smart, 238, 402test cells, 64–6zero error, 411, 413, 416

Insulated bipolar gate transistors (IBGT),199, 201

Intensity: sound, 36–7Interference:

capacitive, 205electrical signals, 199–202electromagnetic, 199–202, 206–7inductive, 204–5measurement interference, 199–202

Interlocks: CVS systems, 346Internal combustion (IC) engines:

connecting rods, 22degrees of freedom, 22emissions chemistry, 325–6energy balance, 268–75friction loss distribution, 360principle axes, 22

Interpretation: specifications, 7–8In-vehicle data transfer protocols, 401–2IP see Institute of Petroleum standardsIrradiation: wear measurement, 361–2ISO 3046 standard, 316–17ISO 8178 standard, 334–5Isolated foundation blocks, 33

JFET transistors, 296

434 Index

Keyless connections, 179–80Keyways, 179–80Kinematic viscosity: oil, 358‘Knocking’: detonation, 284, 303

Lambda ratio, 284, 285Large-speed diesel engines, 56–7Latent heat of evaporation, 96Layouts:

cables, 203–7storage plans, 130test cell plans, 64–6

Lean burn engines, 328Legionnaires’ disease, 99Legislation:

classifications, 332–6emissions, 324, 329–36, 347, 350environmental, 324, 329

Length: supply lines, 207Level of probability, 417, 418Light commercial vehicles, 332–3Light duty:

certification, 331classification, 332exhaust emission test procedure,

348–9Lighting: test cells, 61Limited pressure cycle, 267Liquefied natural gas (LNG), 134, 264

see also Natural gasLiquefied petroleum gas (LPG), 265Liquid fuels, 242–7, 264Lloyd’s Rule Book, 176, 316LNG see Liquefied natural gasLoad calculations, 79–80Load cells, 144, 146Loading vehicles, 386Local issues, 4–5Losses see Heat losses; Mechanical loss

measurementLow level emissions testing site choice,

350–1Low level inlet ducting, 84LPG see Liquefied petroleum gasLubricating oils see OilLubrication testing, 354–67

acronyms, 354–5lubricant certification/classification,

354–5

reference lubricants, 358test regimes, 359–62

Lucas–Dawe air mass flowmeter, 260LVDT transducers, 323

Mains power see Power supplyManagement, 308–23

data, 404–5health and safety, 308, 309–12project roles, 9–10project tools, 11–12risk, 309–12task allocation, 310

Manometers, 235Mapping engines, 322, 395, 400–1Marine engines, 272, 334–5Mass flowmeters, 245–7Mass fraction burned, 288, 289, 290–2Massive foundations, 30–4Mass spectrometer, 339Master drawings, 12Materials: absorption coefficients, 40Matting, 33Mean effective pressure, 293–5Measurements:

airbox method, 256–60air consumption, 242, 250, 256–60compression ring oil films, 360–1crankcase blow-by, 249–50displacement, 235–6electrical signal interference, 199–202exhaust temperatures, 409–12fuel consumption, 242–62gaseous emissions, 337–43gas flows, 250heat losses, 263–81instrumentation, 232, 242–7interference, 199–202liquid fuel consumption, 242–7mechanical losses, 263–81noise, 37–8oil consumption, 242–62particulate emissions, 336pressure, 233–41standards, 317temperature, 236–7, 409–12vehicle noise, 40vibration, 236wear by irradiation, 361–2

Index 435

Mechanical efficiency, 271, 276, 356Mechanical loss measurement, 263–81Medium pressure hot water (MPHW)

supply, 74Medium-speed diesel engines, 56–7Membrane couplings, 182Meters:

blow-by, 250consumption rate, 245–7see also Flowmeters

Methanol, 363Microfog water systems, 68–9Middle market, 302–3Mileage accumulation dynamometers,

375–6Mixture preparation: fuel, 142Models, 395, 401, 404Moist air properties, 94–9Moisture content reduction, 97–8MON see Motor octane numberMontreal Protocol, 70Morse test, 278–9MOSFET transistors, 296Motorcycles, 349Motoring test losses, 278–9Motor octane number (MON), 141Mountings:

combined spring/rubber, 32engines, 24–30test beds, 29–30

Movable coolant conditioning units,116–17

MPHW see Medium pressure hot watersupply

Mufflers, 42, 43, 44Multi-bush couplings, 187Multicell laboratories, 19–20

see also Test cellsMulti-cylinder engines, 23Multiple cells, 124–5Multiple disc flexible couplings, 182Multivariate problem handling, 318, 319

Natural gas (NG), 134, 264–5NDIR see Non-dispersive infra-red

analyserNetworks, 204NG see Natural gasNitrogen oxides, 324, 326

Noise, 21, 36–46absorption coefficients, 40exhausts, 42–3external to test facility, 39fans, 89fundamentals, 36–8measurement, 37–8permitted levels, 38–9in test cells, 39–40vehicle measurements, 40weighting curves, 37–8within test cells, 39–40see also Vibration and noise

Noise, vibration and harshness (NVH),368–9, 376–7

‘road shells’, 388vehicles, 40

Nomenclature:control modes, 221–2see also Notation

Nominated alarm tables, 229Non-dispersive infra-red analyser (NDIR),

337Non-electrical starting systems, 166Non-road diesel engines, 334Normal distribution, 418North Sea gas see Natural gasNotation:

air conditioning, 106–7consumption, 261couplings, 195–6electrical design, 213–15exhaust calorimeter, 281exhaust emissions, 352thermal efficiency, 280ventilation, 106–7vibration and noise, 44–5

NOx (oxides of nitrogen), 324, 326Nozzles, 84Number of significant figures, 420NVH see Noise, vibration and harshness

Octane numbers, 141Offloading machines, 385Off road vehicles, 54–5Oil:

characteristics, 357–8compression rings, 360–1consumption measurement, 242, 248–9

436 Index

Oil (cont.)cooling systems, 140dynamic viscosity, 357engines, 140kinematic viscosity, 358storage, 129–43temperature, 115–18, 140total base number, 358

Opacimeters, 336Open plant circuits, 113–14Operating quadrants, 153–4, 161–2Operational envelope specification, 100Operational instrumentation, 217Operational specifications, 2–3, 7Organization:

projects, 1–13test departments, 308–23

Orifices: flow, 256–7, 259Otto cycle, 266, 267–8Outflows: test cells, 15Outlet ducting, 83–5Outline planning permission, 3–4Overspeed protection standards, 317Oxides of nitrogen (NOx), 324, 326Ozone depletion, 69, 70

PAH see Polyaromatic hydrocarbonsParamagnetic detection (PMD)

analyser, 339Parameter relationships, 321PARAMINS manual, 364Particle formation, 326Particulate emissions see EmissionsPasswords, 406Pendulum vibration, 35Performance:

dynamometers, 163–4engines, 247–8, 251

Permanent magnet dynamometers, 159Permitted levels of noise, 38–9Petrol engines see Gasoline enginesPiezoelectric pressure transducers, 296–7Pipes: fuel, 132, 133–4Piston rings, 360–2Pits, 33, 381–4Planning permission, 3–5Planning permits, 4–5Planning regulations, 39Planning test programmes, 312–13

Platinum resistance thermometers (PRTs),236–7

PLCs see Programmable logic controllersPlenums, 84PMD (paramagnetic detection)

analyser, 339Polyaromatic hydrocarbons (PAH), 325Portable stands: engines, 49Positioning machines, 385Position modes, 222–4Positive displacement flowmeters, 261Post-test data processing, 395, 405Power law modes, 223Power output:

air/fuel ratio, 286–7energy balance, 272–3mechanical losses, 276–8

Power supply:connection layouts, 208interconnections, 208–9interference, 207line lengths, 207security, 219specification, 209, 210UK specification, 210uninterrupted, 219

Power test codes, 315, 316–18Powertrain test cells, 53–4Pressure:

absolute, 234atmospheric, 251–2circuits, 296–9combustion, 104–5, 296–9controls, 104–5, 137–8differential, 234engine fuel, 137–8gauges, 408, 410–11manometers, 235measurement, 233–41, 296–9relative, 234transducers, 234–5ventilation system losses, 82, 93

Pressure-crank angle diagram, 288, 289Primary instrumentation, 65Principle axes, 22Production testing, 55–6, 402–3Programmable logic controllers (PLCs),

214, 218, 219–20Projects:

constraints, 5, 9, 11contractors, 8–9

Index 437

control, 12management role, 9–10management tools, 11–12organization, 1–13timing charts, 12–13

Proof design see Feasibility studiesProperties:

air, 250–1diesel fuels, 142, 367gaseous fuels, 265gasoline, 141, 366liquid fuels, 264moist air, 94–9multi-bush couplings, 187oil, 357–8

Protection rating, 200PRTs see Platinum resistance thermometersPsychometry, 94–9Purge systems, 73, 125, 126Pyrometers, 237, 410, 411

Quality: water, 110–11Quasistatic force, 233Quill shafts, 181

Rain excluding tubes, 127Ramp rate, 229Range: units under test, 3Ratios see Air/fuel ratio; Compression ratio‘Raw’ cooling water circuits, 113–15Reactive mufflers see Exhaust resonatorsReduction: moisture content, 97–8Reference fuels/lubricants, 134, 358Regulations:

CVS systems, 344dangerous substances, 129electrical installation, 198explosive atmospheres, 129external noise, 39safety, 67–8test cell safety, 4–5

Relative accuracy, 421Relative humidity, 94Relative pressure relationships, 234Reporting:

post-acquisition data, 405test programmes, 312–14

Research octane number (RON), 141

Residual fuel storage/treatment, 132–3Resilient matting, 33Response times: analysers, 339Responsibility matrix, 11Restraints:

vehicles, 386–7see also Constraints

Results:accuracy, 408–22correlation, 308, 314–15

Rib-decking roof construction, 60Risk assessment, 308, 311Risk management, 309–12Road load equation, 369–71‘Road shells’, 388–9Rolling road dynamometers, 368–94

end-of-line production testing, 373–4genesis, 371–2in-service assessment/tuning, 373see also Chassis dynamometers

Rolls:diameters, 390–1surfaces, 388tyre contact, 390–1

RON see Research octane numberRoofs: test cells, 59–60Rotational speed, 153–4Rubber annular type couplings, 182–3, 189Rubber/spring combination mountings, 32Rules of thumb: energy balance, 16Running away control, 66, 222, 223

SAE standards, 316Safety:

chassis dynamometers, 387–8European standards, 211instrumentation, 217interaction matrix, 212, 217regulations, 4–5, 67–8test cells, 4–5, 216–21ventilation, 73see also Health and safety

Scavenged ducts, 124–5Schenck dynamometers, 145, 147Scoring method: risk, 311Sealed housing for evaporative

determination (SHED), 349–50Secondary instrumentation, 65Secondary sources: vibration and noise, 21

438 Index

Second order forces: engines, 23Security:

data, 406power supply, 219

Seeing engines in test cells, 57–8Seismic blocks, 30, 31, 33Semi-anechoic test cells, 41Sensing errors, 414Sensitive devices, 208–9Sensors, 304Service factors, 186, 188Services:

space, 62–3status displays, 219–20

Shaft-line components, 148Shafts:

couplings, 178–9drive shaft design example, 185–93Goodman diagrams, 179–80guards, 193whirl, 180–1

Shared control rooms, 51, 52SHED (sealed housing for evaporative

determination), 349–50Shelf life: gasoline, 141Shielded cable protection, 201–2Shielded thermocouples, 411Shock loading, 189–90Shut down control, 66, 220Side-by-side stacking: test cells, 52–3Signal chain calibration, 229–30Significant figures: number of, 420Single cylinder diesel engines, 256–7,

291, 294Single roll sets, 380Site choice: test cells, 350–1Six-cylinder engines see Multi-cylinder

enginesSize of test cells, 47–8, 120–1Sluice-gate dynamometers, 145Smart devices/systems, 238, 402Soak areas, 347Software:

choice, 227, 406–7engines, 406–7production test cells, 402–3suppliers, 282

Solids in water, 111Sound intensity, 36–7Spacing of cables, 206–7

Span gases, 340–2Spark ignition engines, 326–8Specifications:

chassis dynamometers, 391–3control/data acquisition systems, 6difficulties, 7energy balance, 17–19interpretation, 7–8power supply, 209, 210suppliers, 6test cells, 17–19, 47–71

Specific enthalpy: air, 96Specific fuel consumption, 286–7Specific humidity: air, 94Speed:

engine governing standards, 317engine maps, 322sensors, 304time intervals, 231–3

Speed and torque modes, 223–5Spillback, 245, 246Spot fans, 86‘Spring and mass’ vibration, 35Spring-mounted seismic blocks, 30, 31Spring/rubber combination mountings, 32Standard cycles, 265–8Standards:

data acquisition, 399–400diesel fuels, 367gasoline, 366ISO, 316–17, 334–5measurements, 317SAE, 316speed governing, 317

Stands:dynamometers, 51, 52engines, 49

Starting engines, 164–6Start-up checks, 220, 221Statement accuracy, 414–16Static: test beds, 54–5Static deflection and natural frequency

relationship, 26Statistical design: experiments, 318–22Status displays: services, 219–20Steady flow energy equations, 268–9Steam generators, 104Stefan–Boltzmann equation, 76Stoichiometric air/fuel ratio, 264, 265, 285,

286, 327Stoichiometric engines, 255, 326

Index 439

Storage:fuel, 129–43layout plans, 130oil, 129–43residual fuels, 132–3

Storing gas distribution system, 340–2Strain gauge accelerometers, 232, 236Strain gauge transducers, 144, 145, 149,

232–5Stress concentrations: couplings,

179–80Subfloor collection: exhaust gases, 126Subfloor construction: test cells, 58Subfloor services space, 85Subsoils, 33Sulphur content: fuels, 328Sumps, 113–14Supplementary cooling fans, 86Suppliers: specifications, 6Supply:

combustion air, 100cooling water, 108fuel, 129–43oil, 129–43

Supply lines see Power supplySupported bedplates, 30–4Support service space, 62–3Suppression systems: fire, 69Surface examination techniques, 357Synchronous dynamometers, 159System integration, 1, 2–3

Tail pipes, 43–4Tandem dynamometers, 160–1Task allocation, 310TBN see Total base numberTemperature:

air, 253–6diesel engine exhausts, 409–12engine fuel, 138–40engine oil, 115–18flooded inlets, 101fuel consumption effects, 139measurement, 236–7, 409–12operational envelope specification, 100PRTs, 236soak areas, 347

Test beds:foundations, 24–9

inclined/static/dynamic, 54–5mountings, 29–30

Test cells:air conditioning, 72ambient conditions guidelines, 199anechoic, 40–2arrangements, 50–3ATEX codes, 67–8automotive engines, 50–3, 55basic minimum, 48–50cell to cell correlation, 314–15chemical energy, 14climatic, 377–80cold testing, 55–6computer role, 228–33connected devices, 228–33control, 64–6, 216–41, 402control room design, 47–71control surface, 15control volume, 14–15cooling water, 108–28

circuit types, 113–20data acquisition, 216–41design, 16, 47–71, 197–215diesel engines, 56–7dimensions, 48diversity factors, 17–19doors, 58–9dynamometers stands, 51, 52electrical design, 197–215emergency stop, 66energy balance, 17–19, 108, 120–1engines

handling systems, 61–2as heat sources, 78heat transfer, 75–7shut down, 66

environment, 199equipment integration, 301exhaust gases, 108, 122–7external noise, 39extraction ducts, 125final specifications, 17–19fire control, 66–71flooring construction, 58fuel supply, 134–6hearing engines, 57–8heat

losses, 79sources, 78–80transfer calculation, 78

440 Index

Test cells (cont.)high end engines, 301–2hot testing, 55, 56inflows, 15instrumentation, 64–6layout planning, 64–6lighting, 61noise within, 39–40outflows, 15output diagrams, 17, 18overall size, 47–8physical environment, 199planning permits, 4–5production

software, 402–3testing, 55–6

purge systems, 125, 126research and development, 53–4roofs, 59–60safety issues, 4–5, 67–8, 216–21seeing engines, 57–8service space, 62–3shared control rooms, 51, 52shut-down procedures, 220side-by-side stacking, 52–3site choice, 350–1size, 47–8, 120–1specifications, 47–71start-up procedures, 220subfloor construction, 58support service space, 62–3thermal issues, 17, 18, 19thermodynamic systems, 14–20typical designs, 48–56ventilation, 72–107

load calculations, 79–80walls, 59–60

Test cycles: emissions legislation, 330–1Test department organization, 308–23Test facility specification, 1–13Test limits, 331Test methods:

diesel fuels, 367gasoline, 366

Test programmes, 312–14, 320, 321Test regimes, 359–62Test sequences:

analysis, 270–1control, 221–7, 228editing, 228

elements, 228–9emissions, 329–31, 346, 347–9energy balance calculation, 275–6mechanical losses, 278–9software choice, 227

Test sheets, 396Test software, 406–7THD see Total harmonic distortionThermal analysis software, 17, 18Thermal cycling tests, 238–9Thermal efficiency, 263–81

gaseous fuels, 264–8internal combustion engines, 268–75

Thermal ratings, 19Thermal shock, 119, 238–9Thermistors, 237Thermocouples, 236, 411Thermodynamic systems: test cells as,

14–20Thermometers, 237Three Mile Island nuclear accident, 219Throttle actuation, 225–6Time intervals, 231–3Timing charts, 12–13Top dead centre determination, 300Torque:

coupling capacity, 190engine maps, 322

Torque flanges, 145–8Torque measurement, 144–69

accelerating conditions, 152–3brushless torque-shafts, 148calibration, 149–52decelerating conditions, 152–3dynamometers, 144–69error assessment, 149–52in-line shafts, 145–8load cells, 144, 146rotational speed, 153–4shaft-line components, 148torque flanges, 145–8trunnion-mounted machines, 144–5

Torque reversal, 189–90Torque-shafts: brushless, 148Torque and speed modes, 223–5Torsion:

coupling characteristics, 172–8,182–3, 191

vibration standards, 317Total base number (TBN), 358Total energy release, 292

Index 441

Total harmonic distortion (THD), 199, 201Traceability, 419Traditional approach: data collection,

395–7Transducers:

boxes, 230chains, 229choice, 231data acquisition, 229displacement measurement, 235–6inductive, 235–6piezoelectric type, 296–7pressure, 234–5, 296–7

Transient engine testing control,239–41

Transmissibility and frequencyrelationship, 28

Trays: cables, 205Treatment:

combustion air, 99–105fuel, 129–43oil, 129–43residual fuels, 132–3see also Air conditioning

Tribology, 354, 355–7bench tests, 356–7engine test examples, 360–1surface examination, 357

Trunking, 205see also Cables

Trunnion-mounted machines, 144–5, 146Turbocharged engines, 90–4, 105, 272,

273–5, 277Turning moments, 173, 174Turnkey contracts, 8Two mass systems, 172Type tests, 315Tyres, 385, 390–1

UDC see Urban driving cycleUK see United KingdomUnbalanced forces, 22Uncertainty, 417–19Uncontrolled stop, 217Underground fuel lines, 133–4Uninterrupted power supply (UPS), 219United Kingdom (UK):

power supply specification, 210standards, 316–17

United States (US):EPA standards, 325, 330, 334, 350,

373, 374exhaust emission test procedure, 348–9federal regulations, CVS systems, 344SAE standards, 316SHED legislation, 350zone classification, 67

Units under test (UUT), 3, 53–4, 230, 408–9Units xv-xviUniversal joints, 181Unmanned running control, 238UPS see Uninterrupted power supplyUrban driving cycle (UDC), 347–8Urban dynamometer driving schedule, 349US see United StatesUUT see Units under test

Values:energy balance, 272–3see also Calorific value

Valves: fuel gauges, 246Vapour detection extinguishing systems, 71Vapour pressure thermometers, 410Variable fill hydraulic machines, 156–7Variable geometry dynamometers, 384–5Variation record sheets, 9–10Vehicles:

climatic test cells, 377–80data transfer protocols, 401–2drag, 371dynamometers, 368–94loading, 386noise measurements, 40restraints, 386–7see also Emissions

Ventilation, 72–107air flow worked example, 91–2component pressure losses, 82control, 87control rooms, 92–4design, 90–4distribution systems, 80–1duct design, 80–1electrical cabinets, 210external ducting systems, 86fans, 83, 87–94heat transfer to air, 79inlet/outlet ducting, 83–5

442 Index

Ventilation (cont.)load calculations, 79–80notation, 106–7pressure losses, 82, 93purge fans, 73safety requirements, 73system layout, 92test cells, 79–80worked example, 90–4

Vibration and noise, 21–46anechoic test cells, 40–2damping ratios, 28frequency and amplitude ratio

relationship, 27fundamentals, 21–4measurement, 236natural frequency relationships, 26notation, 44–5pendulums, 35perception, 30, 31phase relationships, 35secondary sources, 21sources, 21–4‘spring and mass’ vibration, 35static deflection relationships, 26summary, 34–5transmissibility and frequency

relationship, 28see also Noise, vibration and harshness

Viscous flow air meters, 260Volume see Control volumeVolumetric fuel consumption gauges, 243–4

Walls, 59–60, 77–8Water:

compounds in, 112cooling systems, 108–12hardness, 111–12microfog, 68–9quality, 110–11solid inclusions, 111systems, 68–9see also Cooling water...

Water brakes see Hydrokineticdynamometers

Water-cooled friction dynamometers, 160Water-to-water heat exchangers, 119Wear, 357, 361–2

see also Lubrication testingWeight: vehicle classifications, 332Weighting curves, 37–8Wet-bulb temperatures, 96Wheel substitution dynamometers, 381Whirl: shafts, 180–1Willan’s line method, 279Wind tunnels, 378Workflow processes, 297–8

Zero error: instrumentation, 411, 413, 416Zone 2 hazards, 5Zone classification, 67